1911 Encyclopædia Britannica/Palaeobotany

by Clement Reid

PALAEOBOTANY. In the present article the subject of vegetable palaeontology is treated from a botanical point of view. The science of botany is concerned with the vegetable kingdom as a whole, and not merely with the flora now living. The remains of the plants of former periods, which have come down to us in the fossilized state, are almost always fragmentary, and often imperfectly preserved; but their investigation is of the utmost importance to the botanist, as affording the only direct evidence of the past history of vegetable organisms. Since the publication of the Origin of Species the general acceptance of the doctrine of evolution has given a vastly increased significance to palaeontological data. The determination of the course of descent has now become the ultimate problem for the systematist: this is an historical question, and the historical documents available are the remains of the ancient organisms preserved in the rocks. The palaeobotanist thus endeavours to trace the history of plants in the past, with the hope of throwing light on their natural affinities and on the origin of the various groups. His investigations must embrace not only the comparative morphology and anatomy of fossil plants, but also their distribution over the earth’s surface at different periods—a part of the subject which, besides its direct biological interest, has obvious bearings on ancient climatology and geography.

Preservation.—Before considering the results of palaeobotanical research, some account must be given of the way in which the evidence is presented, or, in other words, of the modes of preservation of vegetable remains. These fall under two main heads. On the one hand, there is the mode of preservation which gives rise to casts, moulds and generally impressions, exhibiting the superficial features of the specimen. The great majority of vegetable fossils are of this kind, and the term incrustation is used as a general term to cover all such methods of fossilization. On the other hand, there are specimens in which the tissues of the plant have been permeated by some mineral in solution, which, subsequently setting hard, has fixed and preserved the internal structure, often with astonishing perfection of detail. This second method of fossilization is termed petrifaction. In the case of incrustation the whole substance of the fossilized specimen—e.g., a stem of Sigillaria—may be replaced by mineral matter, such as sandstone or shale, giving a cast of the whole, on the outer surface of which the external markings, such as the bases of leaves and the scars left by their fall, are visible in their natural form. Usually the original organic substance remains as a thin carbonaceous layer forming the surface of the cast, but sometimes it has entirely disappeared. The surrounding matrix will of course show the mould of the cast, with its elevations and depressions reversed. In the case of thin, flat organs such as leaves, the whole organ may be spread out in the plane of stratification, leaving its impress on the overlying and underlying layers. Here there has not necessarily been any replacement of organic by inorganic material; the whole leaf, for example, may remain, though reduced to a carbonaceous film. In such carbonaceous impression not only are the form and markings, such as venation, perfectly preserved, but something of the actual structure may remain. The cuticularized epidermis, especially, is often thus preserved, and may be removed by the use of appropriate reagents and examined microscopically. If sporangia and spores are present they also may persist in a perfectly recognizable form, and in fact much of our knowledge of the fructification of fossil ferns and similar plants has been derived from specimens of this kind.

In many cases internal casts have been formed, some large cavity, such as a fistular pith, having become filled with mineral substance, which has taken the impress of the surrounding structures, such as the wood. The common casts of Calamites are of this nature, representing the form of the hollow medulla, and bearing on their surface the print of the nodal constrictions and of the ridges and furrows on the inner surface of the wood. The whole organic substance may have been removed, or may persist merely as a thin carbonaceous layer. Mistakes have often arisen from confusing these medullary casts with those of the stem as a whole.

Although some information as to minute structure may often be gleaned from the carbonaceous coating of impressions, the fossils preserved by petrifaction are the main source of our knowledge of the structural characters of ancient plants. The chemical bodies which have played the most important part as agents of petrifaction are silicic acid and calcium carbonate, though other substances, such as magnesium carbonate, calcium sulphate and ferric oxide have also been concerned, either as the chief constituents of petrifactions, or mixed with other bodies. A large number of the most important remains of plants with structure preserved are siliceous; this is the case, for example, with the famous French Permo-Carboniferous fossils of St Etienne, Autun, &c., which in the hands of Brongniart, Renault and others have yielded such brilliant scientific results. At a more recent horizon, the silicified specimens of the Mesozoic Gymnosperms from Great Britain, France, and especially North America, are no less important. Calcified specimens are especially characteristic of the British Carboniferous formation; their preservation is equally perfect with that of the silicified fossils, and their investigation by Witham, Binney, Williamson and others has proved no less fertile. In the Coal Measures of England and of certain German and Austrian districts (e.g. Langendreer in Westphalia; Ostrau in Moravia), calcareous nodules, crowded with vegetable fragments of every kind, occur in certain mines embedded in the substance of the coal and representing its raw material in a petrified condition. Even the most delicate tissues, such as cambium and phloem, the endosperm of seeds, or the formative tissue of the growing-point, are frequently preserved cell for cell, both in calcareous and silicious material. As a rule, the petrified remains, all-important for the revelation of structure, are fragmentary, and give little idea of the habit or external characters of the plants from which they were derived. Hence they must be brought into relation with the specimens preserved as casts or impressions, in order to gain a better conception of the plant as a whole. This is often a difficult task, and generally the fragmentary nature of practically all vegetable fossils is the chief hindrance to their investigation. Owing to this, it has become the common practice of palaeobotanists to give distinct generic names to detached parts of plants which may even have belonged to one and the same species. Thus the roots of Sigillaria are called Stigmaria, detached leaves Sigillariophyllum, and the fructifications Sigillariostrobus; the name Sigillaria applies to the stem, which, however, when old and partly decorticated has been called Syringodendron, while its woody cylinder has often been described under the name Diploxylon. This naming of portions of plants, however objectionable, is often not to be avoided; for detached organs constantly have to be described long before their relation to other parts is established which, indeed, may never be accomplished. For example, the form and structure of Stigmaria have long been well known; but it is seldom possible to determine whether a given Stigmaria belonged to Sigillaria, Lepidodendron or some other genus. The correct piecing together of the fragmentary remains is one of the first problems of the palaeobotanist, and the gradual disappearance of superfluous names affords a fair measure of the progress of his science. The recent advance of fossil botany has depended in a very great degree on the study of petrified specimens with their structure preserved; so far, at least, as the older strata are concerned, it is, as a rule, only with the help of specimens showing structure that any safe conclusions as to the affinities of fossil plants can be arrived at.

The subject of coal (q.v.) is treated elsewhere. Here it need only be said that the masses of vegetable substance, more or less carbonized and chemically altered, of which coal is composed, frequently contain cells and fragments of tissue in a condition recognizable under the microscope, as for example spores (sometimes present in great quantities), elements of the wood, fibres of the bark, &c. These remnants, however, though interesting as revealing something of the sources of coal, are too fragmentary and imperfect to be of any botanical importance. In lignite, on the other hand, the organized structure is sometimes excellently preserved. In the Wealden of Belgium, for example, specimens of Ferns and Coniferae occur, in the form of lignite, which can be sectioned, like recent plants, with a razor, and exhibit an almost unaltered structure.

I.—Palaeozoic

The present section is concerned with the botany of the Palaeozoic age, from the oldest rocks in which vegetable remains have been found up to the close of the Permian period. The Glossopteris flora of India and the southern hemisphere, the age of which has been disputed, but is now regarded as for the most part Permo-Carboniferous, is, however, dealt with in the succeeding section, in connexion with the Mesozoic floras. The various groups of plants represented in the Palaeozoic rocks will first be considered in systematic order, after which some account will be given of the succession and distribution of the various floras during the period.

In dealing with the plants of such remote epochs, the relative importance of the various groups, so far as they are known to us, is naturally very different from that which they assume at the present day. There is no evidence that the Angiospermous flowering plants, now the dominant class, existed during the Palaeozoic period; they do not appear till far on in the Mesozoic epoch, and their earlier history is as yet entirely unknown. On the other hand, fern-like seed-plants, known as Pteridosperms, and Gymnosperms belonging almost entirely to families now extinct, were abundant, while the Pteridophyta attained a development exceeding anything that they can now show. Among the lower classes of plants we have scarcely any knowledge of Palaeozoic Bryophyta; Fungi were probably abundant, but their remains give us little information; while, even among the Algae, which are better represented, well characterized specimens are scanty.

With few exceptions, the remains of Palaeozoic Algae are of comparatively little botanical interest. A vast number of “species” have been described, but, as has been said, “by far the greater number of the supposed fossil Algae have no claim to be regarded as authentic records of this class of Thallophytes” (Seward, 1898). The investigations of Nathorst, Williamson Algae. and others have shown that a very large proportion of the casts and impressions attributed to Algae had in all probability a totally different origin. Some represent the tracks or burrows of worms, crustaceans or other animals; others, the course of rills of water on a sandy or muddy shore; others, again, the marks left on the bottom by bodies drifted along by the waves. In cases of doubt, evidence may be obtained from traces of organic structure, from the presence of carbonaceous matter, or, as Zeiller has pointed out, by the remains of animals such as Bryozoa being attached to the cast, showing that it represents a solid body and not a mere cavity or furrow. Evidence from traces of organization is alone conclusive; the presence of carbonaceous matter, though a useful indication, may be deceptive, for the organic substance may have been derived from other sources than the body which left the impression. The mere external form of the supposed Algae is rarely so characteristic as to afford satisfactory evidence of their nature. Some of the better-attested examples, among which are a few of considerable interest, may now be considered. Of Cyanophyceae, as we should expect, the Palaeozoic remains are very doubtful. Gloioconis, found by Renault in a coprolite of Permian age, was regarded by him as a Cyanophycean allied to Gloeocapsa; this may be so, but the argument drawn from the absence of nuclei, considering the extreme rarity of recognizable nuclei even in the best preserved fossil tissues, can hardly be taken seriously. Girvanella, found in Cambrian, Ordovician and Silurian rocks, as well as in later deposits, appears to have played a part in the origination of oolitic rock-structure. It consists of minute interwoven tubular filaments, and has been variously interpreted as possibly representing the sheaths of a Cyanophycean Alga, and as constituting a Siphoneous thallus of the type of the Codieae. The non-cellular order Siphoneae is fairly well represented in Palaeozoic strata, especially by calcareous verticillate forms referable to the family Dasycladeae; the separate tubular joints of the articulated thallus, bearing the prints of the whorled branches, are sometimes cylindrical (Arthroporella, Vermiporella, &c.), sometimes oval (Sycidium) or spherical (Cyclocrinus). These forms, and others like them, go back to the Silurian and Ordovician; while Gyroporella, from the Permian, is another fairly characteristic Siphoneous type. There can be no doubt that the verticillate Siphoneae, a group much isolated among recent organisms, are among the most ancient families of plants. The gigantic Nematophycus, to be described below, has been regarded as having Siphoneous affinities. Little trace of Confervaceae has been found; Confervites chantransioides, apparently consisting of branched cellular filaments, may perhaps represent a Cambrian Confervoid. Cladiscothallus, from the Culm of Russia, in which the filaments are united to form hemispherical or globular tufts, has been compared by Renault to a Chaetophora. This is one of the somewhat doubtful Algae occurring in boghead coal or torbanite, a carbonaceous rock the nature of which has been much disputed, in the law courts as well as in scientific literature. The boghead of Scotland, Autun and New South Wales is regarded by Renault and Bertrand as mainly composed of gelatinous Algae (Pila and Reinschia), having a hollow, saccate thallus formed of a single layer of cells. It may appear surprising that a body containing 65% of carbon should be so largely made up of gelatinous Algae in a comparatively little altered condition, but the material is rich in bitumen, which seems to have replaced the water contained in the organisms when alive. It has recently been stated, however, that the supposed Algae are in reality the megaspores of Vascular Cryptogams. Scarcely anything is known of Palaeozoic Florideae; Solenopora, ranging from the Ordovician to the Jurassic, resembles, in the structure of its thallus, with definite zones of growth. Corallinaceae such as Lithothamnion, and may probably be of the same nature. A branched filamentous organism from the Lower Carboniferous of Scotland, described by Kidston under the name of Bythotrephis worstoniensis, shows some remains of cellular structure, and may probably be a true Alga, resembling some of the filamentous Florideae in habit.

Apart from the multitude of supposed fossil Algae described as “Fucoids” but usually not of Algal nature, and never presenting determinable characters, very little remains that can be referred to Palaeozoic Brown Algae. The most striking of all fossil Algae, however, Nematophycus, may possibly be a Phaeophycean. The first species of the genus, Nematophycus Logani, was discovered by Dawson in 1856 in the Lower and Middle Devonian of Canada, and was described by him as a Conifer under the name of Prototaxites. Carruthers, however, in 1872 established its Algal nature, and gave it the more appropriate name of Nematophycus. In N. Logani the stem, which is found in a silicified state, may be as much as 3 ft. in diameter. The tissue is made up of large, unseptate, occasionally branching tubes, with an undulating vertical course, among which much smaller tubes are irregularly interwoven. Radially placed gaps in the tissue (at first erroneously interpreted as medullary rays, but subsequently more aptly compared to the air-spaces of large Algae) contain very sparse hyphae, which here branch more freely than elsewhere. The concentric rings of growth, which form a characteristic feature, are due to periodic variations in the size of the larger tubes. Transverse septa have occasionally, but rarely, been detected in the smaller hyphae. Penhallow maintains that these smaller tubes arise as branches from the larger, but other observers have failed to confirm this. In N. Storriei, from the Silurian (Wenlock) of South Wales, described by Barber, there is no sharp differentiation of the two kinds of tubes; they are rarely observed to branch, except in the gaps, which in this species are not radially directed. In N. Ortoni (Penhallow), from the Devonian of Canada, the tubes are quite uniform, and there are no spaces or concentric rings. The tubes have their cavity dilated at intervals, and Penhallow has therefore compared them with the trumpet-hyphae of Laminariaceae, but no transverse septa are anywhere visible. Several other species have been described. Carruthers compared the usually non-cellular structure of Nematophycus with that of Siphoneae such as Halimeda, while recognizing the points of resemblance to Laminariaceae (e.g. Lessonia) in the dimensions of the stem and its concentric rings of growth. Later writers, influenced by the occasional occurrence of transverse walls in the smaller hyphae, have laid more stress on Laminariaceous affinities. The existence of these gigantic Algae in Palaeozoic times, attested by such well-preserved specimens, is a fact of great interest, though their systematic position is still an open question. Pachytheca, a spherical organism, usually about the size of a small pea, found in rocks of Silurian and Devonian age, has been much investigated and discussed, without any decisive light having been thrown on its nature. It was once regarded as connected with Nematophycus (with which it sometimes occurs in association), possibly as its fructification. For this view however, there is no evidence, though the tissues of the two fossils are somewhat similar. Pachytheca is formed of cellular filaments resembling those of a Cladophora, irregularly interwoven in the central region, radiating towards the periphery, and often forked. In one case the spherical thallus was found seated in a cup-like receptacle. There can be little doubt of the Algal nature of the fossil, but beyond this it is impossible at present to carry its determination.

On the whole, it cannot be said that the Palaeozoic remains have as yet thrown much light on the evolution of the Algae, though we may not be prepared to maintain, with Zeiller, that plants of this class appear never to have assumed a form very different from that which they present at the present day.

The first evidence for the existence of Palaeozoic Bacteria was obtained in 1879 by Van Tieghem, who found, that in silicified vegetable remains from the Coal Measures of St Étienne the cellulose membranes showed traces of subjection to butyric fermentation, such as is produced at the present day by Bacillus Amylobacter; he also claimed to have detected the organism Bacteria. itself. Since that time a number of fossil Bacteria, mainly from Palaeozoic strata, have been described by Renault, occurring in all kinds of fossilized vegetable and animal debris. The supposed Micrococci present little that is characteristic; the more definite, rod-like form of the Bacilli offers a better means of recognition, though far from an infallible one; in a few cases dark granules, suggestive of endospores, have been found within the rods. On the whole, the occurrence of Bacteria in Palaeozoic times—so probable a priori—may be taken as established, though the attempt to discriminate species among them is probably futile.

Fungi were no doubt abundant among Palaeozoic vegetation. In examining the tissues of fossil plants of that epoch nothing is more common than to meet with mycelial hyphae in and among the cells; in many cases the hyphae are septate, showing that the higher Fungi (Mycomycetes), as distinguished from the more algoid Phycomycetes, already existed. An Fungi. endophytic Fungus referred to the latter group (Peronosporites antiquarius, W. Smith) bears very definite terminal, or intercalary, spherical vesicles, which may probably be regarded as reproductive organs—either oögonia or sporangia. A minute Fungus bearing sporangia, found by Renault in the wood of a Lepidodendron, and named by him Oöchytrium Lepidodendri, is referred with much probability to the Chytridineae. Conceptacles containing Spores, and strongly suggesting the Chytridineous Fungus Urophlyetis, have recently been found, in petrified material, on the leaves of an Alethopteris, which appears to have undergone decay before fossilization set in. Small spores, almost certainly those of Fungi, are very common in the petrified tissues of Palaeozoic plants. Spherical sacs, bearing forked spines, described by Williamson under the name of Zygosporites, are frequent, usually in an isolated state. Professor Seward, however, has found a Zygosporites in situ, terminating an apparently fungal hypha: he suggests a possible comparison with the mould Mucor. Bodies closely resembling the perithecia of Sphaeriaceous Fungi have often been observed on impressions of Palaeozoic plants, and may probably belong to the group indicated. Professor F. E. Weiss has obtained interesting evidence that the symbiotic association between roots and Fungi, known as “Mycorhiza,” already occurred among Carboniferous plants. The few and incomplete data which we at present possess as to Palaeozoic Fungi do not as yet justify any inferences as to the evolution of these plants. The writer is not aware of any evidence for the occurrence of Palaeozoic Lichens.

The important class of the Bryophyta, which, on theoretical grounds, is commonly regarded as more primitive than the Pteridophyta, is as yet scarcely represented among known fossils of Palaeozoic age. In the Lower Carboniferous of Scotland Mr Kidston has found several specimens of a large dichotomous thallus, with a very distinct midrib; Bryophyta. the specimens, referred to the provisional genus Thallites, much resemble the larger thalloid Liverworts. Similar fossils have been described from still older rocks. In one or two cases Palaeozoic plants, resembling the true Mosses in habit, have been discovered; the best example is the Muscites polytrichaceus of Renault and Zeiller, from the Coal Measures of Commentry. In the absence, however, both of reproductive organs and of anatomical structure, it cannot be said that there is at present conclusive evidence for the existence of either Hepaticae or Musci in Palaeozoic times.

Our knowledge of the Vascular Cryptograms of the Palaeozoic period, though recent discoveries have somewhat reduced their relative importance, is still more extensive than of any other class of plants, and in fact it is here that the evidence of Palaeontology first becomes of essential importance to the botanist. They extend back through the Pteridophyta. Devonian, possibly to the Silurian system, but the systematic summary now to be given is based primarily on the rich materials afforded by the Carboniferous and Permian formations, from which our detailed knowledge of Palaeozoic plants has been chiefly derived.

In addition to the three classes, Equisetales, Lycopodiales and Filicales, under which recent Pteridophytes naturally group themselves, a fourth class, Sphenophyllales, existed in Palaeozoic times, clearly related to the Horsetails and more remotely to the Ferns and perhaps the Club-mosses, but with peculiarities of its own demanding an independent position. We further find that, whereas the Ferns of the present day form a well-defined and even isolated class, this was not the case at the time when the primary rocks were deposited. A great group of Palaeozoic fossils, showing evident affinity to Ferns, has proved to consist of seed-bearing plants allied to Gymnosperms, especially Cycads. This important class of plants will be described at the beginning of the Spermophyta under the name Pteridospermeae. The arrangement which we shall adopt for the Palaeozoic Pteridophyta is therefore as follows:—

I.	Equisetales.	II.	Sphenophyllales.
III.	Lycopodiales.	IV.	Filicales.

We must bear in mind that throughout the Palaeozoic period, and indeed far beyond it, vascular plants, so far as the existing evidence shows, were represented only by the Pteridophyta, Pteridosperms and Gymnosperms. Although the history of the Angiosperms may probably go much further back than present records show, there is no reason to suppose that they were present, as such, amongst the Palaeozoic vegetation. Consequently, the Pteridophytes, Gymnosperms and their allies had the field to themselves, so far as regards the higher plants, and filled places in nature which have now for the most part been seized on by families of more modern origin. Hence it is not surprising to find that the early Vascular Cryptograms were, beyond comparison, more varied and more highly organized than their displaced and often degraded successors. It is among the fossils of the Palaeozoic rocks that we first learn the possibilities of Pteridophytic organization.

I. Equisetales.—This class, represented in the recent flora by the single genus Equisetum, with about twenty species, was one of the dominant groups of plants in Carboniferous times. The Calamarieae, now known to have been the chief Palaeozoic representatives of the Horsetail stock, attained the dimensions of trees, reaching, according to Grand’ Eury, a height of from 30 to 60 metres, and showed in all respects a higher and more varied organization than their recent successors.

Their remains occur in three principal forms of preservation. (1) carbonaceous impressions of the leafy branches, the fructifications and other parts; (2) casts of the stem; these are usually internal, or medullary casts, as described above. Around the cast the organic tissues may be represented by a carbonaceous layer, on the outer surface of which the external features, such as the remains of leaves, can sometimes be traced. More usually, however, the carbonaceous film is thin, and merely shows the impress of the medullary cast within; (3) petrified specimens of all parts—stem, roots, leaves and fructifications—showing the internal structure, more or less perfectly preserved. The correlation of these various remains presents considerable difficulties. Casts surrounded by wood, with its structure preserved, have sometimes been found, and have established their true relations. The position of the branches is shown both on casts and in petrified specimens, and has helped in their identification, while the petrified remains sometimes show enough of the external characters to allow of their correlation with impressions. Fructifications have often been found in connexion with leafy shoots, and the anatomical structure of the axis in sterile and fertile specimens has proved a valuable means of identification.

In habit the Calamarieae appear to have borne, on the whole, a general resemblance to the recent Equisetaceae, in spite of their enormously greater bulk. The leaves were constantly in whorls, and were usually of comparatively small size and of simple form. In the oldest known Calamarian, however, Archaeocalamites (Devonian and Lower Carboniferous), the leaves were repeatedly forked. There is evidence that in some, at least, of the Calamarieae the leaves of each verticil were united at the base to form a sheath. The free lamina, however, was always considerably more developed than in the recent family; in form it was usually linear or narrowly lanceolate. Different genera have been founded on leaf-bearing branches of Calamarieae; apart from Archaeocalamites, already mentioned, and Autophyllites (Grand’ Eury), in both of which the leaves were dichotomous, we have Annularia, Asterophyllites and Calamocladus (in Grand’ Eury’s limited sense), with simple leaves. In some species of Annularia the extremely delicate ultimate twigs, bearing whorls of small lanceolate leaves, give a characteristic habit, suggesting that they may have belonged to herbaceous plants; other Annulariae, however, have been traced with certainty into connexion with the stems of large Calamites. In Asterophyllites, the generic distinction of which from Annularia is not always clear, the narrow linear leaves are in crowded whorls, and the ultimate branches distichously arranged; in the Calamocladus of Grand’ Eury—characteristic of the Upper Coal Measures—the whorls are more remote, and the twigs polystichous in arrangement. In all these groups a leaf-sheath has been recognized.

The distribution of the branches on the main stem shows considerable variations, on which genera or sub-genera have been founded by C. E. Weiss. In Archaeocalamites, which certainly deserves generic rank, the branches may occur on every node, but only in certain parts of the stem; the ribs of successive internodes do not alternate, but are continuous, indicating that the leaves were superposed. Using Calamites as a generic name for all those Calamarian stems in which the ribs alternate at the nodes, we have, on Weiss’s system, the following sub-genera: Stylocalamites, branches rare and irregularly arranged; Calamitina, branches in regular verticils, limited to certain nodes, which surmount specially short internodes; Eucalamites, branches present on every node. These distinctions can be recognized on petrified specimens, as well as on the casts, but their taxonomic value is somewhat doubtful. In many Calamites there is evidence that the aerial stem sprang from a horizontal rhizome, as in the common species C. (Stylocalamites) Suckowi; in other specimens the aerial stem has an independent, rooting base.

The anatomical structure of all parts of the plant is now known, in various Calamarieae, thanks more especially to the work of Williamson in England and of Renault in France. The stem has a structure which may be briefly characterized as that of an Equisetum with secondary growth in thickness (fig. 1, Plate). The usually fistular pith is surrounded by a ring of collateral vascular bundle, (see Anatomy of Plants, and Pteridophyta), each of which, with rare exceptions, has an intercellular canal at its inner edge, containing the disorganized spiral tracheae, just as in the recent genus. The cortex is often preserved; in certain cases it was strengthened by hypodermal strands of fibres, as in Equisetum. It is only in the rare cases where a very young twig is preserved that the primary structure of the stem is found unaltered. In all the larger specimens a broad zone of wood, with its elements in radial series, had been added. This secondary wood, in the true Calamites (Arthropitys, Goeppert), has a simple structure comparable to that of the simplest Coniferous woods; it is made up entirely of radial bands of tracheides interspersed with medullary rays. The pitting of the tracheides is more or less scalariform in character, and is limited to the radial walls. In favourable cases remains of the cambium are found on the outer border of the wood, and phloem is also present in the normal position, though it does not seem to have attained any considerable thickness. In the old stems the primary cortex was replaced by periderm, giving rise to a thick mass of bark. The above description applies to the stems of Calamites in the narrower sense (Arthropitys of the French authors), to which the specimens from the British Coal Measures mostly belong. Archaeocalamites appears to have had a similar structure, but in some specimens from the Lower Carboniferous of Burntisland, provisionally named Protocalamites pettycurensis, centripetal wood was present in the stem. In Calamodendron (Upper Coal Measures) the wood has a more complex structure than in Calamites, the principal rays including radial tracts of fibrous tissue, in addition to the usual parenchyma. Arthrodendron (Lower Coal Measures) approaches Calamodendron in this respect. The longitudinal course of the vascular bundles and their relation to the leaves in Calamarieae generally followed the Equisetum type, though more variable and sometimes more complex. The attachment of the branches was immediately above the node, and usually between two foliar traces, as in the recent genus. Where the structure of the leaves is preserved it proves to be of an extremely simple type; the narrow lamina is traversed by a single vascular bundle, separated by a sheath from the surrounding palisade-parenchyma. Stomata of the same structure as in Equisetum have been detected in the epidermis.

Fig. 1.—Calamites. Part of transverse section of a young stem, showing pith, vascular bundles with secondary wood, and cortex. (× about 40.) From a photograph (Scott, “Studies”).		Fig. 4.—Palaeostachya pedunculata. Fertile shoot, bearing numerous cones and a few leaves. After Williamson (Scott, “Studies”).		Plate. Fig. 22.—Lyginodendron oldhamium. Transverse section of stem, showing the pith containing groups of sclerotic cells, the primary xylem-strands, secondary wood and phloem, pericycle and cortex. 𝑙𝑡1-𝑙𝑡5, leaf-traces, numbered according to the phyllotaxis, 𝑙𝑡5 belonging to the lowest leaf of the five; 𝑝ℎ, a group of primary phloem; 𝑝𝑑, periderm, formed from pericycle. (× 3.)
Fig. 5.—Sphenophyllum insigne. Transverse section of stem, showing triangular primary wood, secondary wood, remains of phloem, and primary cortex. (× about 30.) From a photograph (Scott, “Studies”).				A		B
				Fig. 31.—Cordaianthus Penjoni. A, Male catkin in longitudinal section: 𝑎, axis; 𝑏, bracts; 𝑐, 𝑑, filaments of stamens, hearing the pollen-sacs (𝑒 and 𝑓) at the top; 𝑣, apex of axis. (× 61/2.) B, Stamens more highly magnified: 𝑔, vascular bundle of filament; 𝑒, pollen-sac after dehiscence. (× 23.) After Renault (Scott, “Studies”).

The roots (formerly described as a separate genus, Astromyelon) were borne directly on the nodes, not on short lateral branches as in Equisetum. They are of similar structure in all known Calamarieae, the main roots having a large pith, while the rootlets had little or none. The structure is in all respects that typical of roots, as shown by the centripetal primary wood, and the alternation of xylem and phloem groups observable in exceptionally favourable young specimens. A striking feature is the presence of large, radiating intercellular cavities in the cortex, suggesting an aquatic habit. The young roots show a double endodermis, just as in the recent Equisetum.

A considerable number of Calamarian fructifications are known, preserved, some as carbonaceous impressions, others as petrified specimens, exhibiting the internal structure. In many cases the cones have been found in connexion with branches bearing characteristic Calamarian foliage. Almost all strobili of the Calamarieae are constructed on the same general lines as those of Equisetum, with which some agree exactly; in most, however, the organization was more complex, the complexity consisting in the intercalation of whorls of sterile bracts, between those of the sporangiophores. In several cases heterospory, unknown among recent Equisetaceae, has been demonstrated in their Palaeozoic representatives.

Four main types of structure may be distinguished among Calamarian strobili.

Fig. 2—Calamostachys. Diagrammatic longitudinal section of the cone, showing the axis (𝑎𝑥) bearing alternate whorls of bracts (𝑏𝑟) and peltate sporangiophores (𝑠𝑝) with their sporangia (𝑠𝑚). The upturned tips of the bracts are only shown in every alternate verticil.

1. Calamostachys, Schimper. Here the whorls of peltate sporangiophores alternate regularly with those of sterile bracts, the former being inserted on the axis midway between the latter (fig. 2). The sporangiophores, which are usually half as numerous in each verticil as the bracts, have the same form as in Equisetum, but each bears four sporangia only. The spores are frequently found to be still united in tetrads. In some species, e.g. the British C. Binneyana, numerous specimens have been examined and only one kind of spore observed; here, then, there is a strong presumption that the species was homosporous. In other cases, however, e.g. C. Casheana, Will., two kinds of spore occur, in different sporangia, but on the same strobilus and even on the same sporangiophore. The megaspores, of which there are many in the megasporangium, have a diameter about three times that of the microspores. The abortion of certain spores, which is known to have taken place both in the homosporous C. Binneyana and in the megasporangia of C. Casheana, may throw some light on the origin of the heterosporous condition. The bracts were sometimes coherent in their lower part (e.g. C. Binneyana), sometimes free (e.g. C. Ludwigi); in all cases their free extremities formed a protection to the fertile whorl above. In some continental species (e.g. C. Grand’ Euryi, Ren.) radial membranous plates hung down from each verticil of bracts, forming compartments in which the subjacent sporangiophores were enclosed. The anatomy of the axis is essentially similar to that of a young Calamarian twig, with some variations in detail. Strobili of the Calamostachys type occur in connexion both with Annularia and Asterophyllites foliage.

2. Palaeostachya, Weiss. Here, as in the previous genus, sterile and fertile verticils are ranged alternately on the axis of the cone. The main difference is that in Palaeostachya the sporangiophores, instead of standing midway between the whorls of bracts, are inserted immediately above them, springing, as it were, from the axil of the sterile verticil (fig. 3, A). This singular arrangement has suggested doubts as to the correctness of the current interpretation of the Equisetaceous sporangiophore as a modified leaf (cf. Cheirostrobus below). In most other respects the two genera agree; there is evidence for the occurrence of heterospory in some strobili referred to Palaeostachya. The anatomy of the axis is that of a young branch of a Calamite. According to Grand’ Eury, the Palaeostachya fructification was most commonly associated with Asterophyllites foliage. The external aspect of a Palaeostachya is shown in fig. 4 (Plate).

(After Renault. Scott, Studies.)

Fig. 3.

A, Palaeostachya. Diagrammatic longitudinal section of cone, showing the axis (𝑎𝑥) bearing the bracts (𝑏𝑟) with peltate sporangiophores (𝑠𝑝) springing from their axils; 𝑠𝑚, sporangia.

B, Archaeocalamites. Part of cone, showing the axis (𝑎𝑥) bearing peltate sporangiophores (𝑠𝑝) without bracts; 𝑠𝑚, sporangia.

3. Equisetum type of strobilus. In certain cases the strobili of Palaeozoic Calamarieae appear to have had essentially the same organization as in the recent genus, the axis bearing sporangiophores only, without intercalated bracts. It is remarkable that fructifications apparently of this kind have been found by Renault in close association with the most ancient of the Calamarieae—Archaeocalamites. In these strobili the peltate scales, like the vegetative leaves of the plant, are in superposed verticils; each appears to have borne four sporangia (fig. 3, B). Other cones, however, namely, those known as Pothocites, have also been attributed on good grounds to the genus Archaeocalamites; they are long strobili, constricted at intervals, and it is probable that the succession of fertile sporangiophores was interrupted here and there by the intercalation of sterile bracts, which may also have been present, at long intervals, in Renault’s species. Cones from the Middle Coal Measures, described by Kidston under the name of Equisetum Hemingwayi, but probably belonging to one of the Calamarieae, bear a striking external resemblance to those of a recent Equisetum.

4. Cingularia, Weiss. This form of strobilus, from the Coal Measures of Germany, is imperfectly known, and its relation to Calamarieae not beyond doubt. In the lax strobili the sporangiophores, which are not peltate, but strap-shaped, were borne, as C. E. Weiss first showed, immediately below the verticils of bracts, the position thus being the reverse of that in Palaeostachya.

The Palaeozoic Calamarieae, though so far surpassing recent Equisetaceae, both in stature and complexity of organization, clearly belonged to the same class of Vascular Cryptogams. There is no satisfactory evidence for attributing Phanerogamic affinities to any members of the group, and the view, of which Williamson was the chief advocate, that they form a homogeneous Cryptogamic family, is now fully established.

II. Sphenophyllales.—The class of Sphenophyllales, as known to us at present, is of limited extent, embracing the two genera Sphenophyllum and Cheirostrobus, which may serve as types of two families within the class. The characters of Sphenophyllum are known with some completeness, while our knowledge of Cheirostrobus is confined to the fructification; the former will therefore be described first.

1. Sphenophyllum.—The genus Sphenophyllum, of which a number of species have been described, ranging probably from the Middle Devonian, through the Carboniferous, to the Permian or even the Lower Triassic, consisted of herbaceous plants of moderate dimensions. The long, slender stems, somewhat tumid at the nodes, were ribbed, the ribs running continuously through the nodes, a fact correlated with the superposition of the whorled leaves, the number of which in each verticil was some multiple of 3, and usually 6. In the species on which the genus was founded the leaves, as the generic name implies, are cuneate and entire, or toothed on their anterior margin;^[1] in other cases they are deeply divided by dichotomy into narrow segments, or the whorl consists of a larger number (up to 30) of apparently simple, linear leaves, which may represent the segments of a smaller number. The different forms of leaf may occur on the same plant, the deeply divided foliage often characterizing the main stem, while the cuneate leaves were borne on lateral shoots. A comparison, formerly suggested, with the two forms of leaf in Batrachian Ranunculi has not proved to hold good; the idea of an aquatic habit is contradicted by the anatomical structure, and the hypothesis that the plants were of scandent growth is more probable. The species of Sphenophyllum have a graceful appearance, which has been compared with that of the trailing Galiums of hedgerows. Branches sprang from the nodes, though perhaps not truly axillary in position. The cones, more or less sharply differentiated, terminated certain of the branches.

Fig. 6.—Sphenophyllum Dawsoni. Diagram of cone in longitudinal section.
𝑎𝑥,	Axis.
𝑏𝑟,	Bracts.
𝑠𝑝,	Sporangiophores, each bearing a sporangium, 𝑠𝑚.
𝑏𝑟′,	Whorl of bracts in surface view.

The anatomy of the stem of Sphenophyllum, investigated by Renault, Williamson and others, is highly characteristic (fig. 5, Plate). The stem is traversed by a single stele, with solid wood, without pith; the primary xylem is triangular in section, the spiral elements forming one or two groups at each angle, while the phloem occupied the bays, so that the structure resembles that of a triarch root. Two leaf-trace bundles started from each angle of the stele, and forked, in passing through the cortex, to supply the veins of the leaf, or its subdivisions. The cortex was deeply furrowed on its outer surface. The primary structure is only found unaltered in the youngest stems; secondary growth by means of a cambium set in very early, xylem being formed internally and phloem externally in a perfectly normal manner. At the same time a deep-seated periderm arose, by which the primary cortex was soon entirely cut off. The secondary wood in the Lower Carboniferous species, S. insigne, has scalariform tracheides, and is traversed by regular medullary rays, but in the forms from later horizons the tracheides are reticulately pitted, and the rays are for the most part replaced by a network of xylem-parenchyma. There are no recent stems with a structure quite like that of Sphenophyllum; so far as the primary structure is concerned, the nearest approach is among the Psiloteae, with which other characters indicate some affinity; the base of the stem in Psilotum forms some secondary wood. The diarch roots of a Sphenophyllum have been described by Renault, who has also investigated the leaves; they were strongly constructed mechanically, and traversed by slender vascular bundles branching dichotomously.

Fructification.—Williamson thoroughly worked out, in petrified specimens, the organization of a cone which he named Bowmanites Dawsoni; it was subsequently demonstrated by Zeiller that this fructification belonged to a Sphenophyllum, the cones of the well known species S. cuneifolium having a practically identical structure. The type of fructification described by Williamson and now named Sphenophyllum Dawsoni consists of long cylindrical cones, in external habit not unlike those of some Calamarieae. The axis, which in structure resembles the vegetative stem in its primary condition, bears numerous verticils of bracts, those of each verticil being coherent in their lower part, so as to form a disc or cup, from the margin of which the free limbs of the bracts arise. The sporangia, which are about twice as numerous as the bracts, are seated singly on pedicels or sporangiophores springing from the upper surface of the bract-verticil, near its insertion on the axis (fig. 6). As a rule two sporangiophores belong to each bract. The sporangium is attached to the enlarged distal end of its pedicel, from which it hangs down, so as to suggest an anatropous ovule on its funiculus. Dehiscence appears to have taken place at the free end of the sporangium; the spores are numerous, and, so far as observed, of one kind only. Each sporangiophore is traversed throughout its length by a vascular bundle connected with that which supplies the subtending bract. This form of fructification appears, from Zeiller’s researches, to have been common to several species of Sphenophyllum, but others show important differences. Thus Bowmanites Römeri, a fructification fully investigated by Solms-Laubach, differs from S. Dawsoni in the fact that each sporangiophore bears two sporangia, attached to a distal expansion approaching the peltate scale of the Equisetales. It is thus proved that the sporangiophore is not a mere sporangial stalk, but a distinct organ, in all probability representing a ventral lobe of the subtending bract. The recently discovered species, Sphenophyllum fertile, while resembling Bowmanites Römeri in its peltate, bisporangiate sporangiophores, is peculiar in the fact that both dorsal and ventral lobes of the sporophyll were fertile, dividing in a palmate manner into several branches, each of which constitutes a sporangiophore. Thus the sterile bracts of other species are here replaced by sporangium-bearing organs. In Sphenophyllum majus, where the cones are less sharply defined, the forked bract bears a group of four sporangia at the bifurcations, but their mode of insertion has not yet been made out.

Fig. 7.—Cheirostrobus. Diagram of cone, the upper part in transverse, the lower in longitudinal section. In the transverse section six sporophylls, each showing three segments, are represented.
S𝑝.𝑎,	Section through sterile segments.
S𝑝.𝑏,	Section through sporangiophores.
𝑠𝑡,	Laminae of sterile segments.
𝑓,	Peltate expansions of sporangiophores.
𝑠𝑚,	Sporangia.
𝑣.𝑏.	Vascular bundles.
𝑐𝑦,	Stele of axis (A𝑥).
In the longitudinal section the corresponding parts are shown.

2. Cheirostrobeae.—The family Cheirostrobeae is only known from the petrified fructification (Cheirostrobus pettycurensis) derived from the Lower Carboniferous of Burntisland in Scotland. The excellence of the preservation of the specimens has rendered it possible to investigate the complex structure in detail. The cone is of large size—3·5 cm. in diameter; the stout axis bears numerous whorls of compound sporophylls, the members of successive verticils being superposed. The sporophylls, of which there are eleven or twelve in a whorl, are each composed of six segments, three being inferior or dorsal, and three superior or ventral. The dorsal segments are sterile, corresponding to the bracts of Sphenophyllum Dawsoni, while the ventral segments constitute peltate sporangiophores, each bearing four sporangia, just as in a Calamarian fructification (fig. 7). The great length and slender proportions of the segments give the cone a peculiar character, but the relations of position appear to leave no doubt as to the homologies with the fructification of Sphenophylleae; as regards the sporangiophores, Bowmanites Römeri occupies exactly the middle place between S. Dawsoni and Cheirostrobus. The axis of the cone in Cheirostrobus contains a polyarch stele, with solid wood, from the angles of which vascular bundles pass out, dividing in the cortex, to supply the various segments of the sporophylls. In the peduncle of the strobilus secondary tissues are formed. While the anatomy has a somewhat Lycopodiaceous character, the arrangement of the appendages is altogether that of the Sphenophylleae; at the same time Calamarian affinities are indicated by the characters of the sporangiophores and sporangia.

The Sphenophyllales as a whole are best regarded as a synthetic group, combining certain characters of the Ferns and Lycopods with those of the Equisetales, while showing marked peculiarities of their own. Among existing plants their nearest affinities would appear to be with Psiloteae, as indicated not merely by the anatomy, but much more strongly by the way in which the sporangia are borne. There is good reason to believe that the ventral synangium of the Psiloteae corresponds to the ventral sporangiophore with its sporangia in the Sphenophyllales. Professor Thomas of Auckland, New Zealand, has brought forward some interesting variations in Tmesipteris which appear to afford additional support to this view.

Pseudobornia.—Professor Nathorst has described a remarkable Devonian plant, Pseudobornia ursina (from Bear Island, in the Arctic Ocean), which shows affinity both with the Equisetales and Sphenophyllales. The stem is articulated and branched, attaining a diameter of about 10 cm. The smaller branches bear the whorled leaves, probably four in each verticil. The leaves are highly compound, dividing dichotomously into several leaflets, each of which is deeply pinnatifid, with fine segments. When found detached these leaves were taken for the fronds of a Fern. The fructification consists of long, lax spikes, with whorled sporophylls; indications of megaspores have been detected in the sporangia. The discoverer makes this plant the type of a new class, the Pseudoborniales. At present only the external characters are known.

(After Stur. Scott, Studies.)
Fig. 8.—Leaf-base of a Lepido­dendron.
𝑠.𝑐.,	Scar left by the leaf.
𝑣.𝑏.,	Print of vascular bundle.
𝑝, 𝑝,	Parichnos.
𝑙,	Ligule.
𝑎, 𝑎,	Superficial prints below scar.

III. Lycopodiales.—In Palaeozoic ages the Lycopods formed one of the dominant groups of plants, remarkable alike for the number of species and for the great stature which many of them attained. The best known of the Palaeozoic Lycopods were trees, reaching 100 ft. or more in height, but side by side with these gigantic representatives of the class, small herbaceous Club-mosses, resembling those of the present day, also occurred. Broadly speaking, the Palaeozoic Lycopods, whatever their dimensions, show a general agreement in habit and structure with our living forms, though often attaining a much higher grade of organization. We will first take the arborescent Lycopods, as in every respect the more important group. They may all be classed under the one family Lepidodendreae, which is here taken to include Sigillaria.

Lepidodendreae.—The genus Lepidodendron, with very numerous species, ranging from the Devonian to the Permian, consisted of trees, with a tall upright shaft, bearing a dense crown of dichotomous branches, clothed with simple narrow leaves, ranged in some complex spiral phyllotaxis. In some cases the foliage is preserved in situ; more often, however, especially in the main stem and larger branches, the leaves had been shed, leaving behind them their scars and persistent bases, on which the characteristic sculpturing of the Lepidodendroid surface depends. The cones, often of large size, were either terminal on the smaller twigs, or, it is alleged, borne laterally on special branches of considerable dimensions. At its base the main stem terminated in dichotomous roots or rhizophores, bearing numerous rootlets. To these underground organs the name Stigmaria is applied; they are not clearly distinguishable from the corresponding parts of Sigillaria. The numerous described species of Lepidodendron are founded on the peculiarities of the leaf-cushions and scars, as shown on casts or impressions of the stem. The usually crowded leaf-cushions are spirally arranged, and present no obvious orthostichies, thus differing from those of Sigillaria. Each leaf-cushion is slightly prominent; towards its upper end is the diamond-shaped or triangular scar left by the fall of the actual leaf (fig. 8). On the scar are three prints, the central one alone representing the vascular bundle, while the lateral prints (parichnos) mark the position of merely parenchymatous strands. In the median line, immediately above the leaf-scar, is a print representing the ligule, or rather the pit in which it was seated. On the flanks of the cushion, below the scar, are two superficial prints, perhaps comparable to lenticels. In the genus Lepidophloios the leaf-cushions are more prominent than in Lepidodendron, and their greatest diameter is in the transverse direction; on the older stems the leaf-scar lies towards the lower side of the cushion. The genus Bothrodendron, going back to the Upper Devonian, differs from Lepidodendron in its minute leaf-scars and the absence of leaf-cushions, the scars being flush with the smooth surface of the stem. In the Lower Carboniferous of central Russia beds of coal occur consisting of the cuticles of a Bothrodendron, which are not fossilized, but retain the consistency and chemical composition of similar tissues in recent plants.

(Scott, Studies.)
Fig. 9.—Lepidodendron Veltheimianum. Transverse section of stem.
𝑝,	Pith, almost destroyed.	𝑝ℎ,	Phloem and pericycle.
𝑥,	Zone of primary wood.	𝑏𝑟,	Stele of a branch.
𝑝𝑥,	Protoxylem.	𝑝𝑑,	Periderm.
𝑥2,	Secondary wood.	𝑙.𝑏,	Leaf-bases.
The primary cortex between stele and periderm has perished. (✕41/2.)

The anatomy of Lepidodendron and its immediate allies is now well known in a number of species; the Carboniferous rocks of Great Britain are especially rich in petrified specimens, which formed the subject of Williamson’s extensive investigations. The stem is in all cases monostelic; in most of the forms the central cylinder underwent secondary growth, and the distinction between primary and secondary wood is very sharply marked. In L. Harcourtii, however, the species earliest investigated (by Witham, 1833, and Brongniart, 1837), and in one or two other species, no secondary wood has yet been found. The primary wood of Lepidodendron forms a continuous cylinder, not broken up into distinct bundles; its development was clearly centripetal, the spiral elements forming more or less prominent peripheral groups. In the larger stems of most species there was a central pith, but in certain of the smaller branches, and throughout the stem in some species (L. rhodumnense, L. selaginoides), the wood was solid. A single leaf-trace, usually collateral in structure, passed out into each leaf. The primary structure of the stem was thus of a simple Lycopodiaceous type, resembling on a larger scale what we find in the upright stem of Selaginella spinosa. In most species (e.g. L. selaginoides, L. Wunschianum, L. Veltheimianum) secondary growth in thickness took place, and secondary wood was added, in the centrifugal direction, showing a regular radial arrangement, with medullary rays between the series of tracheides (fig. 9). The tissue thus formed often attained a considerable thickness. While primary phloem can be recognized with certainty in favourable cases, the question of the formation of secondary phloem by the cambium is not yet fully cleared up. In the Lepidodendron fuliginosum of Williamson, shown by its leaf-bases to have been a Lepidophloios, the secondary wood is very irregular, and consists largely of parenchyma. The same is the case in Lepidodendron obovatum, one of the few species in which both external and internal characters are known. The occurrence of secondary growth in these plants, demonstrated by Williamson’s researches, is a point of great interest. Some analogy among recent Lycopods is afforded by the stem of Isoëtes, and by the base of the stem in Selaginella spinosa; in the fossils the process was of a more normal type, but some of its details need further investigation. The cortex, often sharply differentiated into sclerotic and parenchymatous zones, is bordered externally by the persistent leaf-bases. The development of periderm was a constant feature, and this tissue attained a great thickness, consisting chiefly of a phelloderm, produced on the inner side of the formative layer, and no doubt subserving a mechanical function.

The structure of a Bothrodendron has recently been investigated and proves to be identical with that of the petrified stem which Williamson named Lepidodendron mundum. The anatomy is of the usual medullate Lepidodendroid type; no secondary growth has yet been detected in the stem.

The most interesting point in the structure of the leaf-base is the presence of a ligule, like that of Isoëtes or Selaginella, which was seated in a deep pit, opening on the upper surface of the cushion, just above the insertion of the lamina. The latter shows marked xerophytic adaptations; the single vascular bundle was surrounded by a sheath of short tracheides, and the stomata were sheltered in two deep furrows of the lower surface.

Fig. 10.—Lepidostrobus. Diagram of  cone, in longitudinal section.
2𝑎𝑥,	Axis, bearing the sporophylls (𝑠𝑝ℎ), on each of which a sporangium (𝑠𝑚) is seated.
𝑙𝑔,	Ligule.
The upper sporangia contain numerous microspores; in each of the lower sporangia four megaspores are shown.

The cones of Lepidodendron and its immediate allies are for the most part grouped under the name Lepidostrobus. These cones, varying from an inch to a foot in length, according to the species, were borne either on the ordinary twigs, or, as was conjectured, on the special branches (Ulodendron and Halonia) above referred to. In Ulodendron the large circular, distichously arranged prints were supposed to have been formed by the pressure of the bases of sessile cones, though this interpretation of the scars is open to doubt, and it is now more probable that they bore deciduous vegetative branches; in the Halonial branches characteristic of the genus Lepidophloios the tubercles may perhaps mark the points of insertion of pedunculate strobili. The organization of Lepidostrobus is essentially that of a Lycopodiaceous cone. The axis, which in anatomical structure resembles a vegetative twig, bears numerous spirally arranged sporophylls, each of which carries a single large sporangium on its upper surface (fig. 10). The sporophyll, usually almost horizontal in position, has an upturned lamina beyond the sporangium, and a shorter dorsal lobe, so that the form of the whole is somewhat peltate. A ligule is present immediately below the lamina, its position showing that the whole of the elongated horizontal pedicel on which the sporangium is seated corresponds to the short base of a vegetative leaf. The sporangia, usually of very large size compared with those of most recent Lycopods, have a palisade-like outer wall, and contain either an immense number of minute spores or a very small number of exceedingly large spores (fig. 10). It is very doubtful whether any homosporous Lepidostrobi existed, but there is reason to believe that here, as in the closely allied Lepidocarpon, microsporangia and megasporangia were in some cases borne on different strobili. In other species (e.g. in the cone attributed to the Lower Carboniferous Lepidodendron Veltheimianum) the arrangement was that usual in Selaginella, the microsporangia occurring above and the megasporangia below in the same strobilus (diagram, fig. 10). The genus Spencerites (Lower Coal Measures) differs from Lepidostrobus mainly in the insertion of the sporangium, which, instead of being attached along the whole upper surface of the sporophyll, was connected with an outgrowth on its upper surface by a small neck of tissue towards the distal end. The spores of this genus are curiously winged, and intermediate in size between the microspores and megaspores of Lepidostrobus; the question of homospory or heterospory is not yet decided. The cones of Bothrodendron and another form named Mesostrobus are in some respects intermediate between Lepidostrobus and Spencerites. A more important deviation from ordinary Lepidostroboid structure is shown by the genus Lepidocarpon, from the English Coal Measures and the Lower Carboniferous of Scotland. In this fructification the organization is at first altogether that of a Lepidostrobus; in each megasporangium, however, only a single megaspore came to maturity, occupying almost the whole of the sporangial cavity (see fig. 12), but accompanied by the remains of its three abortive sister cells. An integument grew up from the superior surface of the sporophyll, completely enveloping the sporangium, except for a narrow crevice left open along the top. In favourable cases the prothallus is found preserved, within the functional megaspore or embryo-sac, and the whole appearance, especially as seen in a section tangential tangential to the strobilus, is then remarkably seed-like (see diagram, fig. 11). The seed-like body was detached as a whole from the cone, and in this condition was known for many years under the name of Cardiocarpon anomalum, having been wrongly identified with a true Gymnospermous seed so named by Carruthers. The analogies with a seed are obvious; the chief difference is in the micropyle, which is not tubular, but forms a long crevice, running in a direction radial to the strobilus. Lepidocarpon affords a striking instance of homoplastic modification, for there is no reason to suppose that the Lycopods were on the line of descent of any existing Spermophyta. In a male cone, probably belonging to Lepidocarpon Lomaxi, the microsporangia are provided with incomplete integuments.

Fig. 11.—Lepidocarpon Lomaxii. Diagrammatic section of “seed” in plane tangential to the parent strobilus.
𝑠𝑝ℎ,	Sporophyll.
𝑣𝑏,	Its vascular bundle.
𝑖,	Integument.
𝑚,	Micropylar crevice.
𝑎,	Base.
𝑠𝑚,	Wall of sporangium.
𝑚𝑔,	Membrane of functional megaspore, which is filled by the prothallus, 𝑝𝑟.

Another case of a “seed-bearing” Lycopod has lately been discovered by Miss Benson in Miadesmia membranacea, a slender Selaginella-like plant from the Lower Coal Measures of Lancashire. The female fructification is in the form of a rather lax strobilus. Each sporophyll bears a megasporangium, attached to its upper surface at the proximal end, containing a single large megaspore (fig. 13). The megasporangium is enclosed in an integument, which completely envelopes it, leaving only a narrow micropyle at the distal end (fig. 13). The long tentacles of the integument may have served to facilitate pollination. The seed-like character of the organ is even more striking in Miadesmia than in Lepidocarpon. There seems to be no near affinity between these genera, in which the seed-habit must have arisen independently.

(Scott, Studies.)
Fig. 12.—Lepidocarpon Lomaxii. Sporangium and sporophyll before development of integument. (✕ about 12.)
𝑐𝑢,	Lateral cushions on sporophyll.
𝑣𝑏,	Vascular bundle,
𝑤𝑝,	Palisade layer of sporangium-wall.
𝑤𝑖,	Inner layer of wall.
𝑎,	Base of sporangium.
𝑚𝑔,	Membrane of mega­spore or embryo-sac.

Sigillaria.—The great genus Sigillaria, even richer in “species” than Lepidodendron, ranges throughout the Carboniferous, but has not yet been detected in earlier rocks. The Sigillariae, like the Lepidodendra, were large trees, but must have differed from those of the previous group in habit, for they appear to have branched sparingly or not at all, the lofty upright shaft terminating, like some modern Xanthorrhaea, in a great sheaf of long, grass-like leaves. The strobili were stalked, and borne on the main stem, among the leaves. The roots, or at least their functional representatives, resembled those of Lepidodendron. The chief distinctive character of Sigillaria lies in the arrangement of the leaf-scars, which form conspicuous vertical series on the surface of the stem. In one great division of the genus—the Eusigillariae—the stems are ribbed, each rib bearing a vertical row of leaf-scars; the ribbed Sigillariae were formerly divided into two sub-genera—Rhytidolepis, with the scars on each rib rather widely spaced, and Favularia, where they are approximated and separated by transverse furrows, each rib thus consisting of a series of contiguous leaf-bases. This distinction, however, has proved to have no constant taxonomic value, for both arrangements may occur on different parts of the same specimen. The species without ribs—Subsigillariae—were in like manner grouped under the two sub-genera Clathraria and Leiodermaria; in the former each scar is seated on a prominent cushion, while in the latter the surface of the stem (as in Bothrodendron) is perfectly smooth. Here also the distinction has proved not to hold good, S. Brardi, for example, showing both conditions on the same stem. All these names, however, are still in use as descriptive terms. Generally, the Eusigillariae are characteristic of the older Carboniferous strata, the Subsigillariae of the Upper Coal Measures and Permian. The leaf scars throughout the genus show essentially the same prints as in Lepidodendron, differing only in details, and here also a ligule was present (fig. 14).

(From a drawing by Mrs D. H. Scott. Scott, Studies.)
Fig. 13.—Miadesmia membranacea. Radial longitudinal section of seed-like organ. (✕ about 30.)
𝑙,	Lamina of sporophyll.	𝑙𝑔,	Ligules.
𝑣𝑏,	Vascular bundle.	𝑠𝑚,	Sporangium-wall.
𝑣,	Velum or integument.	𝑚,	Membrane of megaspore.
𝑡,	Tentacles.

(After Weiss. Scott, Studies.)
Fig. 14.—Sigillaria Brardi. Part of surface of stem, showing five leaf-scars. (✕ 11/2.)
𝑣𝑏,	Print of vascular bundle.
𝑝𝑎,	Parichnos.
𝑙𝑔,	Ligule.

The anatomy of Sigillaria is not so well known as that of Lepidodendron, for specimens showing structure are comparatively rare, a fact which may be correlated with the infrequency of branching in the genus. The structure of a Clathrarian Sigillaria (S. Menardi), from the Permian of Autun, was accurately described by Brongniart as long ago as 1839, and a similar species, S. spinulosa (＝S. Brardi) was investigated by Renault in 1875, but it was long before we had any trustworthy data for the anatomy of the ribbed forms. This gap in our knowledge has now been filled up, owing to Bertrand’s investigation of a specimen referred by him to S. elongata, followed by the detailed researches of Kidston and Arber on Sigillaria elegans, scutellata and mamillaris. The structure of the ribbed Sigillariae, as at present known, essentially resembles that of a medullate Lepidodendron, though the ring of primary wood is narrower. Its outer margin is crenulated, the leaf-traces being given off from the middle of each bay. Secondary wood was formed in abundance, precisely as in most species of Lepidodendron. In the Subsigillarian species S. Menardi the primary wood is broken up into distinct bundles, while in S. spinulosa their separation is sometimes incomplete. The secondary cortex or periderm attained a great development, and in some cases shows considerable differentiation. On the whole, the anatomy of Sigillaria is closely related to that of the preceding group, and in fact a continuous series can be traced from the anatomically simplest species of Lepidodendron to the most modified Sigillariae. The leaves of Sigillaria are in some cases almost identical in structure with those of Lepidodendron, but in certain species (S. scutellata and S. mamillaris) there is evidence that they were of the Sigillariopsis type, the leaf being traversed by two parallel vascular strands, derived from the bifurcation of the leaf-trace.

The nature of the fructification of Sigillaria was first satisfactorily determined in 1884 by Zeiller, who found the characteristic Sigillarian leaf-scars on the peduncles of certain large strobili (Sigillariostrobus). The cones, of which several species have been described, bear a strong general resemblance to Lepidostrobus, differing somewhat in the form of the sporophylls and some other details. The megaspores (reaching 2 mm. or more in diameter) were found lying loose on the sporophylls by Zeiller; the sporangia containing them were first observed by Kidston, in a species from the Coal Measures of Yorkshire. That the cones were heterosporous there can be no doubt, though little is known as yet of the microsporangia. The discovery of Sigillariostrobus, which was the fructification of Subsigillariae as well as of the ribbed species, has finally determined the question of the affinities of the genus, once keenly discussed; Sigillaria is now clearly proved to have been a genus of heterosporous Lycopods, with the closest affinities to Lepidodendron.

Stigmaria.—On present evidence there is no satisfactory distinction to be drawn between the subterranean organs of Sigillaria and those of Lepidodendron and its immediate allies, though some progress in the identification of special forms of Stigmaria has recently been made. These organs, to which the name Stigmaria was given by Brongniart, have been found in connexion with the upright stems both of Sigillaria and Lepidodendron. In the Coal Measures they commonly occur in the underclay beneath the coal-seams. Complete specimens of the stumps show that from the base of the aerial stem four Stigmarian branches were given off, which took a horizontal or obliquely descending course, forking at least twice. These main Stigmarian axes may be 2 to 3 ft. in diameter at the base, and 30 or 40 ft. in length. Their surface is studded with the characteristic scars of their appendages or rootlets, which radiated in all directions into the mud. Petrified specimens of the main Stigmaria are frequent, and those of its rootlets extraordinarily abundant. The two parts are very different in structure: in the main axis, as shown in the common Coal Measure form Stigmaria ficoides, the centre was occupied by the pith, which was surrounded by a zone of wood, centrifugally developed throughout. In other species, however, the centripetal primary xylem is represented. Phloem, surrounding the wood, is recognizable in good specimens; in the cortex the main feature is the great development of periderm. The rootlets, which branched by dichotomy, contain a slender monarch stele exactly like that in the roots of Isoëtes and some Selaginellae at the present day; they possessed, however, a complex absorptive apparatus, consisting of lateral strands of xylem, connecting the stele with tracheal plates in the outer cortex. The morphology of Stigmaria has been much discussed; possibly the main axes, which do not agree perfectly either with rhizomes or roots, may best be regarded as comparable with the rhizophores of Selaginellae; they have also been compared with the embryonic stem, or protocorm, of certain species of Lycopodium; the homologies of the appendages with the roots of recent Lycopods appear manifest. It has been maintained by some palaeobotanists that the aerial stems of Sigillaria arose as buds on a creeping rhizome, but the evidence for this conclusion is as yet unconvincing.

Lycopoditeae.—Under this name are included the fossil Lycopods of herbaceous habit, which occur occasionally, from the Devonian onwards. One such plant, Miadesmia, has already been referred to, as one of the seed-bearing Lycopods. In some Lycopoditeae the leaves were all of one kind, while others were heterophyllous, like most species of Selaginella. The genus Selaginellites, Zeiller, is now used to include those forms in which the fructification has proved to be heterosporous. In Selaginellites Suissei there was a definite strobilus bearing both micro- and megasporangia; in each of the latter from 16 to 24 megaspores were contained; in Selaginellites primaevus, however, the number of megaspores was only 4, and the resemblance to a recent Selaginella was thus complete. Selaginellites elongatus, another heterosporous species, is remarkable for having no differentiated strobilus, a condition not known in the recent genus. The antiquity of the Selaginella type indicates that this group had no direct connexion with the Lepidodendreae, but sprang from a distinct and equally ancient herbaceous stock. There is, however, some evidence that Isoëtes, which in several respects agrees more nearly with the Lepidodendreae, may actually represent their last degenerate survivors (see Pleuromeia, in § II., Palaeobotany). No homosporous Lycopoditeae have as yet been recognized.

IV. Filicales.—Of all Vascular Cryptogams the Ferns have best maintained their position down to the present day. Until recently it has been supposed that the class was well represented in the Palaeozoic period, and, indeed, that it was relatively, and perhaps absolutely far richer in species even than in the recent flora. Within the last few years, however, the position has completely changed, and the majority of the supposed Palaeozoic Ferns are now commonly regarded as more probably seed-bearing plants, a conclusion for which, in certain cases, there is already convincing evidence. The great majority of specimens of fossil fern-like plants are preserved in the form of carbonaceous impressions of fronds, often of remarkable perfection and beauty. The characters shown by such specimens, however, when, as is usually the case, they are in the barren state, are notoriously unstable, or of small taxonomic value, among recent plants. Hence palaeobotanists have found it necessary to adopt a purely artificial system of classification, based on form and venation of the frond, in the absence of adequate data for a more natural grouping. The well-known form-genera Pecopteris, Sphenopteris, Odontopteris, &c., are of this provisional nature. The majority of these fronds have now fallen under suspicion and can no longer be accepted as those of Ferns; the indications often point to their having belonged to fern-like Spermophyta, as will be shown below.

It has thus become very difficult to decide what Palaeozoic plants should still be referred to the Filices. The fructifications by themselves are not necessarily decisive, for in certain cases the supposed sporangia of Marattiaceous Ferns have turned out to be in reality the microsporangia or pollen-sacs of seed-bearing plants (Pteridosperms). It is, however, probable that a considerable group of true Ferns, allied to Marattiaceae, existed in Palaeozoic times, side by side with simpler forms. In one respect the fronds of many Palaeozoic Ferns and Pteridosperms were peculiar, namely, in the presence on their rachis, and at the base of their pinnae, of anomalous leaflets, often totally different in form and venation from the ordinary pinnules. These curious appendages (Aphlebiae), at first regarded as parasitic growths, have been compared with the feathery outgrowths which occur on the rachis in the Cyatheaceous genus Hemitelia, and with the anomalous pinnules found in certain species of Gleichenia, at the points of bifurcation of the frond.

(After various authors. Scott, Studies.)

Fig. 15.—Group of Palaeozoic fructifications of Ferns or Pteridosperms.

A, Asterotheca. 1, Pinnule bearing 8 synangia. 2, Synangium in side view. 3, In section, magnified.

B, Renaultia. 1, Fertile pinnule, nat. size. 2, Sporangium, enlarged.

C, Dactylotheca, as in B.

D, Sturiella. Section of pinnule and synangium. 𝑎, Vascular bundle; 𝑐, hairs; 𝑏, 𝑑, annulus, magnified.

E, Oligocarpia. Sorus in surface-view, magnified.

F, Crossotheca. Fertile pinnule, bearing several tufts of microsporangia, magnified.

G, Senftenbergia. Group of annulate sporangia, magnified.

H, Hawlea. Synangium after dehiscence, magnified.

J, Urnatopteris. 1, Part of fertile pinna, nat. size. 2, Sporangia, showing apical pores, magnified.

Of the above. A, D, E, G and H, probably belong to true Ferns; F is the male fructification of a Pteridosperm (Lyginodendron); the rest are of doubtful nature.

Marattiaceae.—A considerable number of the Palaeozoic fern-like plants show indications—more or less decisive—of Marattiaceous affinities; some account of this group will first be given. The reference of these ferns to the family Marattiaceae, so restricted in the recent flora, rests, of course, primarily on evidence drawn from the fructifications. Typically Marattiaceous sori, consisting of exannulate sporangia united to form synangia, are frequent, and are almost always found on fronds with the character of Pecopteris, large, repeatedly pinnate leaves, resembling those of Cyatheaceae or some species of Nephrodium. In certain cases the anatomical structure of these leaves is known, and found to agree generally with that of recent coriaceous fern-fronds. The petiole was usually traversed by a single vascular bundle, hippocrepiform in section—a marked point of difference from the more complex petioles of recent Marattiaceae. There is evidence that in many cases these Pecopteroid fronds belonged to arborescent plants, the stems on which they were borne reaching a height of as much as 60 ft. These stems, known as Megaphytum when the leaves were in two rows, and as Caulopteris in the case of polystichous arrangement, are frequent, especially in the Permian of the Continent; when petrified, so that their internal structure is preserved, the name Psaronius is employed. The structure is often a complex one, the central region containing an elaborate system of numerous anastomosing steles, accompanied by sclerenchyma; the cortex is permeated or coated by a multitude of adventitious roots, forming a thick envelope to the stem. The whole structure bears a general resemblance to that of recent Marattiaceae, though differing in detail. We will now describe some of the fructifications, which are grouped under generic names of their own; these genera, as having a more natural basis, tend to supersede the artificial groups founded on vegetative characters. The genus Asterotheca includes a number of Ferns, chiefly of Coal Measure age, with fronds of the Pecopteris type. The sori, or synangia, ranged in two series on the under-side of the fertile pinnules, are circular, each consisting of 3 to 6 sporangia, attached to a central receptacle and partly united to each other (fig. 15, A); the sporangia separated when mature, dehiscing by a ventral slit. Stur’s genus Hawlea (fig. 15, H), characterized by the separation of the sporangia, may only represent an advanced stage of an Asterotheca. In Ptychocarpus the fusion of the sporangia to form the synangium was much more complete; Scolecopteris resembles Asterotheca, but each synangium is stalked. In all these genera there is an obvious similarity to the synangia of Kaulfussia, while in some respects Marattia or Danaea is approached. In another Pecopteroid genus, Sturiella, the synangia resemble those of Asterotheca, but each sporangium is provided with a band of enlarged cells of the nature of an annulus (fig. 15, D). As a similar differentiation, though less marked, appears in the recent genus Angiopteris, the presumption is in favour of the Marattiaceous affinities of Sturiella, which also shows some relation to the genus Corynepteris (see below, Botryopterideae). In the genus Danaeites, from the Coal Measures of the Saar, the synangia are much like those of the recent Danaea, each sporangium opening by an apical pore. In the Grand' Eurya of Stur the sporangia appear to have been free from each other, as in Angiopteris. On the whole there is thus good evidence for the frequency of Marattiaceae in the Palaeozoic period, though the possibility that the fructifications may really represent the microsporangia of fern-like spermophytes must always be borne in mind. In a certain number of genera the reference to Marattiaceae is much more doubtful. In Dactylotheca, for example (fig. 15, C), a Pecopteroid genus, ranging throughout the Carboniferous, the elongated sporangia individually resemble those of Marattiaceae, but they are completely isolated, the characteristic grouping in sori being absent; the same remark applies to the Sphenopteroid Renaultia of Zeiller (fig. 15, B); the foliage of Sphenopteris, one of the most extensive of Palaeozoic frond-genera, with many different types of fructification, resembled that of various species of Asplenium or Davallia. In many fern-like plants of this period the fronds were dimorphic, the fertile leaves or pinnae having a form quite different from that of the vegetative portions. This was the case in Urnatopteris (Kidston), with Sphenopteroid sterile foliage; the sporangia, borne on the filiform pinnules of the fertile rachis, appear to have dehisced by an apical pore (fig. 15, J). The magnificent Devonian Fern Archaeopteris hibernica, with a somewhat Adiantiform habit, bore special fertile pinnae; the fructification is still imperfectly understood, but the presence of stipules, observed by Kidston, has been adduced in support of Marattiaceous affinities. In all these cases there is reason to suspect that the plants may have been Pteridosperms, rather than Ferns.

Other Families.—The Marattiaceae are the only recent family of Ferns which can be supposed to have existed in anything like its present form in Palaeozoic times. Of other recent orders the indications are meagre and dubious, and there can be no doubt that a large proportion of Ferns from the older rocks (in so far as they were Ferns at all) belonged to families quite distinct from any which we recognize in the flora of our own day. Little or nothing is known of Palaeozoic Ophioglossaceae. Certain fructifications have been referred to Gleicheniaceae (Oligocarpia, fig. 15, E), Schizaeaceae (Senftenbergia, fig. 15, G), Hymenophyllaceae and Osmundaceae, and on good grounds, so far as the external characters of the sporangia are concerned; our knowledge of most of the Ferns in question is, however, far too incomplete to justify us in asserting that they actually belonged to the families indicated. In the case of the Osmundaceae there is good evidence, from anatomical characters, for tracing the family back to the Palaeozoic; their oldest members show a distinct relationship to the Botryopterideae, described in the next paragraph. Numerous more or less isolated fern-sporangia occur in the petrified material of the Carboniferous formation; the presence of an annulus is a frequent character among these specimens, while synangic sori are rare; it is thus certain that families remote from the Marattiaceae were abundantly represented during this period.

(After Renault.)

Fig. 16.—Zygopteris pinnata.

A, Group of sporangia, in surface view.

B, Single sporangium, in transverse section, showing annulus on both sides, magnified.

Botryopterideae.—The family Botryopterideae, first discovered by Renault, stands out with striking clearness among the Palaeozoic Ferns, and differs widely from any group now in existence. The Botryopterideae are chiefly known from petrified specimens; in the genus Botryopteris and certain species of Zygopteris we have a fairly complete knowledge of all parts of the plant. The type-genus Botryopteris, represented in the Permo-Carboniferous of France and in both the Lower and Upper Carboniferous of Great Britain, had a rhizome, with a very simple monostelic structure, bearing spirally arranged compound leaves, with lobed pinnules, probably of a somewhat fleshy texture. In the French species, B. forensis, the plant covered with characteristic jointed hairs, which have served to identify the various organs on which they occur. The sporangia were large pyriform sacs, shortly stalked, and borne in tufts on the branches of the fertile rachis, which developed no lamina. Each sporangium had, on one side only, a longitudinal or slightly oblique annulus, several cells in width; the numerous spores were all of the same size; certain differences among them, which have been interpreted as indicating heterospory, have now proved to depend merely on the state of preservation. The genus Zygopteris, of which numerous Carboniferous and Permian species are known, likewise had a monostelic stem, but the structure of its vascular cylinder was somewhat complex, resembling that of the most highly differentiated Hymenophyllaceae, with which some species of Zygopteris also agreed in the presence of axillary shoots. There is evidence that the stem in some species was a climbing one; the pinnate leaves, arranged on the stem in a two-fifths spiral, were dimorphic, the sterile fronds resembling some forms of Sphenopteris. The petioles have a somewhat complex structure, the bundle often having, in transverse section, the form of an H; it has been proposed to subdivide the genus on the details of the petiolar structure. It is characteristic of Zygopteris and its near allies that two rows of pinnae were borne on each side of the rachis, at least in the fertile fronds. On the fertile rachis the sporangia were borne in tufts, much as in the preceding genus; they were still larger, reaching 2.5 mm. in length, and had a multiseriate annulus, extending, however, to both sides of the sporangium (see fig. 16, A and B). In Stauropteris, a genus showing some affinity with Zygopteris, the branched rachis of the fertile frond terminates in fine branchlets, each bearing a single, spherical sporangium, without any differentiated annulus (fig. 17). The spores in the sporangia have been found in a germinating condition; the stages of germination correspond closely with those observed in recent homosporous ferns (fig. 18). This fact strongly confirms the conclusion, drawn from morphological and anatomical characters, that the Botryopterideae were true Ferns. The genus Corynepteris of Baily is interesting from the fact that its sporangia, while individually similar to those of Zygopteris, were grouped in sori or synangia, resembling those of an Asterotheca. The family Botryopterideae appears to have included a number of other genera, though in most cases the evidence from vegetative structure is alone available. The genus Diplolabis of Renault, shows much in common with Zygopteris as regards anatomical structure, but resembles Corynepteris in possessing a synangic fructification. The genus Asterochlaena of Corda with a deeply-lobed stele, goes back to the Devonian. The family as a whole is of great interest, as presenting points of contact with various recent orders, especially Hymenophyllaceae, Osmundaceae and Ophioglossaceae; the group appears to have been a synthetic one, belonging to a primitive stock (the Primofilices of Arber) from which the later Fern families may have sprung.

(From a drawing by Mrs D. H. Scott. Scott, Studies.)

Fig. 17.—Stauropteris oldhamia. Three sporangia borne on branchlets of the rachis. In A the stomium (𝑠𝑡) or place of dehiscence is shown. B is cut tangentially. In C, 𝑝 is the palisade tissue of the rachis. (✕ about 35.)

A number of genera of Palaeozoic “fern-fronds” have been described, of the fructification of which nothing is known. This is the case, for example, with Diplotmema, a genus only differing from Sphenopteris in the dichotomy of the primary pinnae, and with Mariopteris, which bears a similar relation to Pecopteris. The same holds good of the Pecopteroid Ferns included under Callipteris and Callipteridium. In such cases, as will be explained below, there is a strong presumption that the fronds were not those of Ferns, but of seed-bearing plants of the new class Pteridospermeae.

(From a drawing by Mr L. A. Boodle. Scott, Studies.)

Fig. 18.—Stauropteris oldhamia. Four germinating spores from the interior of a sporangium. All four are putting out rhizoids. In C, lying horizontally, an additional cell has been cut off between rhizoid and spore. (✕ 335.)

On the present evidence it appears that the class Filicales was well represented in the Palaeozoic flora, though by no means so dominant as was formerly supposed. The simpler Ferns (Primofilices) of the period are for the most part referred to the remarkable family Botryopterideae, a group very distinct from any of the more modern families, though showing analogies with them in various directions. On the other hand there was the far more complex Marattiaceous type, strikingly similar in both vegetative and reproductive characters to the recent members of the family. Although doubts have lately been cast on the authenticity of Palaeozoic Marattiaceae owing to the difficulty in distinguishing between their fructifications and the pollen-bearing organs of Pteridosperms, the anatomical evidence (stem of Psaronius) strongly confirms the opinion that a considerable group of these Ferns existed.

Spermophyta.—The Pteridospermeae, for which Potonié’s name Cycadofilices is still sometimes used, include all the fern-like plants which, on the evidence available, appear to have been reproduced by means of seeds. The cases in which such evidence is decisive are but few, namely, Lyginodendron oldhamium, Neuropteris heterophylla, Pecopteris Pluckeneti, Pteridospermeae. Aneimites fertilis and Aneimites tenuifolius. In the first-named plant the structure, both of the vegetative and reproductive organs, is known, and the evidence, from comparison and association, is sufficiently strong. In the other cases there is direct proof of continuity between seed and plant, but only the external characters are known. In a great number of forms, amounting to a majority of the Palaeozoic plants of fern-like habit, the indirect evidence is in favour of their having possessed seeds. We will begin with the Lyginodendreae, a group in which the anatomical characters indicated a systematic position between Ferns and Cycads, long before the reproductive organs were discovered.

(After Williamson. Scott, Studies.)
Fig. 19.—Heterangium Grievii. Restoration of Stem, shown partly in transverse and longitudinal section, partly in surface view.
𝑥,	Primary wood.
𝑥2,	Secondary wood.
𝑝.𝑐.	Phloem and pericycle.
𝑐,	Cortex.
ℎ𝑦,	Hypoderma.
𝑙.𝑡, 𝑙.𝑡,	Leaf-traces.
𝑟,	Adventitious root. Several leaf-bases are shown.

(After Stur. Scott, Studies.)

Fig. 20.—Sphenopteris elegans (foliage of Heterangium Grievii). Part of frond. (2/3 nat. size.)

(Scott, Studies.)
Fig. 21.—Heterangium Grievii. Part of the stele of the stem in transverse section, showing a primary xylem-strand and adjacent tissues (✕ 135.)
𝑝𝑥,	Protoxylem of strand.
𝑥,	Centripetal.
𝑥1,	Centrifugal primary wood.
𝑚𝑥,	Part of the internal wood.
𝑐.𝑝.	Conjunctive tissue.
𝑥2,	Secondary wood.
𝑐𝑏,	Cambium.
𝑝ℎ2,	Phloem.

Lyginodendreae.—Of the genus Heterangium, which still stands very near the true Ferns, several species are known, the oldest being H. Grievii, of Williamson, from the Lower Carboniferous of Scotland. This plant had a long, somewhat slender, ridged stem, the ridges corresponding to the decurrent bases of the spirally arranged leaves (fig. 19). The specimens on which the genus was founded are petrified, showing structure rather than habit, but conclusive evidence has now been obtained that the foliage of H. Grievii was of the type of Sphenopteris (Diplotmema) elegans (fig. 20), and was thus in appearance altogether that of a Fern, with somewhat the habit of an Asplenium. The stem has a single stele, resembling in general primary structure that of one of the simpler species of Gleichenia; there is no pith, the wood extending to the centre of the stele. The leaf-traces, where they traverse the cortex, have the structure of the foliar bundles in Cycads, for they are of the collateral type, and their xylem is mesarch, the spiral elements lying in the interior of the ligneous strand. The leaf-traces can be distinguished as distinct strands at the periphery of the stele, as shown in fig. 21. Most of the specimens had formed a zone of secondary wood and phloem resembling the corresponding tissues in a recent Cycad; the similarity extended to minute histological details, as is shown especially in H. tiliaeoides, a Coal Measures species, where the preservation is remarkably perfect. The cortex was strongly constructed mechanically; in addition to the strands of fibres at the periphery, horizontal plates of stone-cells were present in the inner cortex, giving both stem and petiole a transversely striated appearance, which has served to identify the different parts of the plant, even in the carbonized condition (cf. figs. 19 and 20). The single vascular bundle which traversed the petiole and its branches was concentric, the leaves resembling those of Ferns in structure as well as in habit. Heterangium shows, on the whole, a decided preponderance of Filicinean vegetative characters, though in the leaf-traces and the secondary tissues the Cycads are approached. The organs of reproduction are not yet known, though there is a probability that an associated seed allied to Lagenostoma (see below) belonged to Heterangium. In the Coal Measure genus Megaloxylon, of Seward, which in structure bears a general resemblance to Heterangium, the primary wood consists for the most part of short wide tracheides; probably, as the secondary tissues increased, it had become superfluous for conducting purposes, and was adapted rather for water-storage. In the genus Lyginodendron, of which L. oldhamium, from the Coal Measures, is now the best-known of all Palaeozoic plants, the central wood has disappeared altogether and is replaced by pith; the primary wood is only represented in the leaf-trace strands, which form a ring of distinct collateral bundles around the pith; thus the “medullate-monostelic” structure characteristic of the higher plants was already attained. The individual bundles, however, have the same structure as in Heterangium, and agree closely with the foliar bundles of Cycads. The secondary tissues, which are highly developed, are also of a Cycadean character (fig. 22, Plate). The vegetative organs of the plant are very completely known; the foliage has proved to be that of a Sphenopteris, identical with the species long known under the name of S. Höninghausi, Apart from the important advance shown in the anatomy of the stem, Lyginodendron agrees structurally with Heterangium. There is reason to believe that Lyginodendron oldhamium was a climbing plant comparable in some respects to such recent Ferns as Davallia aculeata. The roots were at first like those of Marattiaceae but grew in thickness like the roots of Gymnosperms.

(From a model after Oliver.)

Fig. 23.—Lagenostoma Lomaxii (the seed of Lyginodendron). Restoration of a seed, enclosed in the lobed cupule, which bears numerous glands. (✕ about 15.)

   (From a photograph. Scott, Studies.)

Fig. 24.—Capitate gland on the cupule of Lagenostoma Lomaxii. (✕ 70.)

The first definite evidence of the mode of reproduction of Lyginodendron oldhamium was due to F. W. Oliver, who in 1903 identified the seed, Lagenostoma Lomaxii, by means of the glands on its cupule, which agree exactly with those on the associated leaves and stems of the plant (cf. figs. 24 and 25). No similar glands are known on any other Palaeozoic plant. Lagenostoma Lomaxii is a small barrel-shaped seed (5·5 by 4·25 mm. when mature) enclosed in a husk or cupule, which completely enveloped it when young, but was ultimately open (figs. 23 and 26 and fig. 27 from another species). The seed was stalked, and there is an exact agreement in structure between the vascular strands of the stalk and cupule of the seed, and those of the rachis and leaflets of Lyginodendron, thus confirming the evidence from the glands. The seed itself is of a Cycadean type, and radially symmetrical. The single integument is united to the nucellus, except at the top, and is traversed by about nine vascular strands. In the apex of the nucellus, as in most Palaeozoic seeds and in recent Cycads, a pollen-chamber, for the reception of the pollen-grains or microspores, is excavated (fig. 26). In Lagenostoma the pollen-chamber has a peculiar structure, a solid column of tissue rising up in the middle, leaving only a narrow annular crevice, in which pollen-grains are found. The neck of the flask-shaped pollen-chamber projected a little from the micropyle and no doubt received the pollen directly. The seed, which need not be described in further detail, was a highly organized structure, showing little trace of the cryptogamic megasporangium from which we must suppose it to have been derived. From the structure of the seed-bearing stalk, and from the analogy of the similar form Lagenostoma Sinclairi (fig. 27) it appears that the seed was borne on a leaf, or part of a leaf, reduced to a branched rachis.

   (From a photograph. Scott, Studies.)

Fig. 25.—Capitate Gland on the Petiole of Lyginodendron oldhamium. (✕ 70.)

The male organs of Lyginodendron were discovered by Kidston, a year or two after the seeds were identified. They are of the type known as Crossotheca, formerly regarded as a Marattiaceous fructification. The genus is characterized by the arrangement of the sporangia, which hang down from the lower surface of the little oval fertile leaflets, the whole resembling an epaulet with its fringe (fig. 15, F; fig. 28). In the case of Lyginodendron the Crossotheca occurs in connexion with the vegetative parts of the frond. Each fertile pinnule bore six, or rarely seven fusiform microsporangia, described as bilocular; not improbably each may represent a synangium. The microspores are tetrahedral. This is the first case in which the pollen-bearing organs of a Pteridosperm have been identified with certainty.

It will be seen that, while the seeds of Lyginodendron were of an advanced Cycadean type, the microsporangiate organs were more like those of a Fern, the reproductive organs thus showing the same combination of characters which appears in the vegetative structure. The family Calamopityeae, allied anatomically to Lyginodendreae, is of Devonian and Lower Carboniferous age.

A,	Micropylar region.
B,	Body of seed.
C,	Chalazal region.
D,	Stalk.
𝑐,	Cupule, surrounding seed.
𝑣𝑏,	Vascular bundles of stalk, cupule and integument.
𝑐𝑝,	“Canopy,” or water-reservoir, at top of integument.
𝑝𝑐,	Cavity of pollen chamber.
𝑐𝑐,	Central column.
𝑎𝑝𝑐,	Aperture of pollen chamber.

(After Oliver. Scott, Studies.)

Fig. 26.—Lagenostoma Lomaxii. Diagram of seed in median longitudinal section.

(After Arber. Scott, Studies.)

Fig. 27.— Lagenostoma Sinclairi. Two seeds, enclosed in lobed cupules and borne on branches of the rachis. (✕ 5.)

Cycadoxyleae.—A few Coal Measure and Permian stems (Cycadoxylon and Ptychoxylon) resemble Lyginodendron in the general character of their tissues, but show a marked reduction of the primary wood, together with an extensive development of anomalous wood and bast around the pith, a peculiarity which appears as an individual variation in some specimens of Lyginodendron oldhamium. It is probable that these stems belonged to plants with the fructification and foliage of Cycads, taking that group in the widest sense. It is only quite at the close of the Palaeozoic period that Cycads begin to appear. The Lyginodendreae type of structure, however, appears to have formed the transition not only to the Cycadales, but also to the extinct family Cordaiteae, the characteristic Palaeozoic Gymnosperms (see p. 107).

(From a sketch after Kidston. Scott, Studies.)

Fig. 28.—Crossotheca Höning­hausi, the male fructification of Lyginodendron. Fertile leaflets, bearing sporangia, and sterile leaflets on the rachis of the same leaf. (✕ 2.)

Medulloseae.—In some respects the most remarkable family of the Cycad-fern alliance is that of the Medulloseae, seed-bearing plants often of great size, with a fern-like foliage, and a singularly complex anatomical structure without parallel among recent plants. Some of the Medulloseae must have had a habit not unlike that of tree-ferns, with compound leaves of enormous dimensions, belonging to various frond-genera—especially, as has now been proved, to Alethopteris and Neuropteris; these are among the most abundant of the Carboniferous fronds commonly attributed to Ferns, and extend back to the Devonian. In habit some species of Alethopteris resembled the recent Angiopteris, while the Neuropteris foliage may be compared with that of an Osmunda. The Medullosa stems have been found chiefly in the Permo-Carboniferous of France and Germany, but a Coal Measures species (M. anglica) has been discovered in Lancashire. The great anatomical characteristic of the stem of the Medulloseae is its polystelic structure with secondary development of wood and bast around each stele. In M. anglica, the simplest species known, the steles are uniform, and usually only three in number; the structure of the stem is essentially that of a polystelic Heterangium. In the Permo-Carboniferous species, such as M. stellata and M. Leuckarti, the arrangement is more complicated, the steles showing a differentiation into a central and a peripheral system; the secondary growth was extensive and unequal, usually attaining its maximum on the outer side of the peripheral steles. In certain cases the structure was further complicated by the appearance of extrafascicular zones exterior to the whole stelar system. The spirally arranged petioles (Myeloxylon) were of great size, and their decurrent bases clothed the surface of the stem; their structure is closely similar to that of recent Cycadean petioles; in fact, the leaves generally, like those of Stangeria at the present day, while fern-like in habit, were Cycadean in structure. In the case of Medullosa anglica we have an almost complete knowledge of the vegetative organs—stem, leaf and root; Cycadean characters no doubt predominate, but the primary organization of the stem was that of a polystelic Fern. In the new genus Sutcliffia, also from the Coal Measures of Lancashire, the stem had a single, large central stele, from which smaller strands were given off, forming a kind of network, which gave rise to the numerous concentric leaf-traces which entered the petioles. This plant may be regarded as anatomically the most primitive of the Medulloseae.

(After Kidston. Scott, Studies.)

Fig. 29.—Neuropteris heterophylla. Seed, attached to a branch of the rachis bearing two vegetative leaflets. (✕ 2.)

In one member of the Medulloseae, there is direct evidence of reproduction by seeds, for in Neuropteris heterophylla Kidston has demonstrated that large seeds, of the size of a hazel-nut, were borne on the frond (fig. 29). In this case the internal structure is not known, but another seed, Trigonocarpus Parkinsoni, associated with, and probably belonging to, the Alethopterid species, Medullosa anglica, occurs in the petrified condition and has been fully investigated. This is a large seed, with a very long micropyle; it has a beaked pollen-chamber, and a complex integument made up of hard and fleshy layers, closely resembling the seed of a modern Cycad; the nucellus, however, was free from the integument, each having its own vascular system. Various other seeds of the same type are known, and in a great number of instances Grand’ Eury has found the fronds of Neuropterideae (Medulloseae) in close association with definite species of seeds, so there can be little doubt that the whole family was seed-bearing. Very little is known at present of the male organs. Some authors have been so much impressed by the similarity of this extinct family to the Cycads, that they have regarded them as being on the direct line of descent of the latter group; it is more probable, however, that they formed a short divergent phylum, distinct, though not remote, from the Cycadean stock.

Pecopterideae.—It has now been established that the form-genus Pecopteris, once regarded as representing the typical Marattiaceous foliage, was in part made up of seed-bearing plants. In 1905 Grand’ Eury discovered the seeds of Pecopteris Pluckeneti, an Upper Coal Measure species, attached, in immense numbers, to the fronds, which are but little modified as compared with the ordinary vegetative foliage. The seeds are flat and winged, closely resembling those of some Cordaiteae (see below). Another form of fructification, compared to the sori of Dicksonia, appears to represent the male organs. There is reason to believe that other species of Pecopteris and similar genera, (Callipteris and Mariopteris) bore seeds, though the artificial group Pecopterideae probably also includes the fronds of true Marattiaceous Ferns.

Aneimiteae.—The genus Aneimites, resembling the Maidenhair Ferns in habit, has now been transferred to the Pteridosperms, the seeds having been discovered in 1904 by David White. In A. fertilis, from the Pottsville beds (Millstone Grit) of West Virginia, the rhomboidal seeds, flattened and winged like those of Cordaiteae, are borne terminally on the lateral pinnae of a frond, which elsewhere bears the characteristic cuneiform leaflets. Continuity between seeds and frond was also demonstrated in another species, A. tenuifolius. The allied genus Eremopteris occurs in association with seeds of a similar platyspermic type.

The Pteridosperms, of which only a few examples have been considered, evidently constituted a group of vast extent in Palaeozoic times. In a large majority of the Fern-like fossils of that period the evidence is in favour of reproduction by seeds, rather than by the cryptogamic methods of the true Ferns. The class, though clearly allied to the typical Gymnosperms, may be kept distinct for the present on account of the relatively primitive characters shown in the anatomy and morphology, and may be provisionally defined as follows: plants resembling Ferns in habit and in many anatomical characters, but bearing seeds of a Cycadean type; seeds and microsporangia borne on fronds only slightly modified as compared with the vegetative leaves.

Gymnospermous remains are common in Palaeozoic strata from the Devonian onwards. The investigations of the last quarter of the 19th century established that these early representatives of the class did not, as a rule, belong to any of its existing families, but formed for the most part a distinct group, that of the Cordaitales, which has Gymnosperms. long since died out. Specimens of true Cycads or Conifers are rare or doubtful until we come to the latest Palaeozoic rocks. Our knowledge of the Cordaiteae (the typical family of the class Cordaitales) is chiefly due to the French investigators, Grand’ Eury and Renault, who successfully brought into connexion the various fragmentary remains, and made known their exact structure.

Cordaitales.—The discovery of the fossil trunks and of their rooted bases has shown that the Cordaiteae were large trees, reaching 30 metres or more in height; the lofty shaft bore a dense crown of branches, clothed with long simple leaves, spirally arranged. Fig. 30, founded on one of Grand’ Eury’s restorations, gives an idea of the habit of a tree of the genus Dorycordaites, characterized by its lanceolate acute leaves; in the typical Cordaites they were of a blunter shape, while in Poacordaites they were narrow and grass-like. The leaves as a rule far exceeded in size those of any of the Coniferae, attaining in some species a length of a metre. Of living genera, Agathis (to which the Kauri Pine of New Zealand belongs) probably comes nearest to the extinct family in habit, though at a long interval. The stem resembled that of Cycads in having a large pith, sometimes as much as 4 in. in diameter; the wood, however, was dense, and had the structure of that of an Araucarian Conifer; specimens of the wood have accordingly been commonly referred to the genus Araucarioxylon, and at one time the idea prevailed that wood of this type indicated actual affinity with Araucarieae. Other characters, however, prove that the Cordaiteae were remote from that family, and the name Araucarioxylon is best limited to wood from later horizons, where a near relationship to Araucarieae is more probable.^[2] In some cases the external tissues of the Cordaitean stem are well preserved; the cortex possessed a system of hypodermal strands of fibres, comparable to those found in the Lyginodendreae. In most cases the leaf-traces passed out from the stem in pairs, as in the recent Ginkgo; dividing up further as they entered the leaf-base. In many Cordaiteae the pith was discoid, i.e. fistular and partitioned by frequent diaphragms, as in some species of Pinus and other plants at the present day. The curious, transversely-ribbed fossils known as Sternbergia or Artisia have proved to be casts of the medullary cavity of Cordaiteae; their true nature was first demonstrated by Williamson in 1850. In those stems which have been referred with certainty to the Cordaiteae there is no centripetal wood; the spiral elements are adjacent to the pith, as in a recent Conifer or Cycad; certain stems, however, are known which connect this type of structure with that of the Lyginodendreae; this, for example, is the case in the Permian genus Poroxylon, investigated by Bertrand and Renault, which in general structure has much in common with Cordaiteae, but possesses strands of primary wood, mainly centripetal, at the boundary of the pith, as in the case in Lyginodendron. Stems (Mesoxylon) intermediate in structure between Poroxylon and Cordaites have lately been discovered in the English Coal Measures. Corresponding strands of primary xylem have been observed in stems of the genus Pitys (Witham), of Lower Carboniferous age, which consisted of large trees, probably closely allied to Cordaites. There appears, in fact, so far as stem-structure is concerned, to have been no sharp break between the typical Palaeozoic Gymnosperms and pronounced Pteridosperms such as Lyginodendron.

(After Grand’ Eury, modified. Scott, Studies.)

 Fig. 30.—Dorycordaites. Restoration, showing roots, trunk and branches bearing long lanceolate leaves and fructifications. The trunk is shown too short.

The long, parallel-veined leaves of the Cordaiteae, which were commonly referred to Monocotyledons before their structure or connexion with other parts of the plant was known, have been shown by Renault to have essentially the same anatomy as a single leaflet of a Cycad such as Zamia. The vascular bundles, in particular, show precisely the characteristic collateral mesarch or exarch structure which is so constant in the recent family (see Anatomy of Plants). In fact, if the foliage alone were taken into account, the Cordaiteae might be described as simple-leaved Cycads. The reproductive organs, however, show that the two groups were in reality very distinct. Both male and female inflorescences have frequently been found in connexion with leaf-bearing branches (see restoration, fig. 30). The inflorescence is usually a spike bearing lateral cones or catkins, arranged sometimes distichously, sometimes in a spiral order. The investigation of silicified specimens has, in the hands of Renault, yielded striking results. A longitudinal section of a male Cordaianthus (the name applied to isolated fructifications) is shown in fig. 31, A, Plate. The organ figured is one of the catkins (about a centimetre in length) which were borne laterally on the spike. Some of the stamens are inserted between the bracts, in an apparently axillary position, while others are grouped about the apex of the axis. Each stamen consists of a long filament, bearing several erect, cylindrical pollen-sacs at its summit (cf. fig. 31, B, Plate). Some of the pollen-sacs had dehisced, while others still retained their pollen. The stamens are probably best compared with those of Ginkgo, but they have also been interpreted as corresponding to the male “flowers” of the Gnetaceae. In any case the morphology of the male Cordaitean fructification is clearly very remote from that of any of the Cycads or true Coniferae, though some resemblance to the stamens of Araucarieae may be traced. The female inflorescences vary considerably in organization; in some species the axis of the spike bears solitary ovules, each accompanied by a few bracts, while in others the lateral appendages are catkins, each containing from two to several ovules. In the catkin shown in longitudinal section in fig. 32, A, it appears that each ovule was borne terminally, on an extremely short axillary shoot, as in Taxus among recent Gymnosperms. The ovule consists of an integument (regarded by some writers as double) enclosing the nucellus. In the upper part of the nucellus is a cavity or pollen-chamber, with a narrow canal leading into it, precisely as in the ovules of Stangeria or other Cycads at the present day (fig. 32, B). Within the pollen-chamber, and in the canal, pollen-grains are found, agreeing with those in the anthers, but usually of larger size (fig. 32, C). It was in this case that Renault first made the exceedingly interesting discovery that each pollen-grain contains a group of cells, presumably representing an antheridium (fig. 32, C). Recent observations have completely confirmed Renault’s interpretation of the facts, on which some doubt had been cast. In the isolated seeds of Cordaitales and Pteridosperms, pollen-grains are often found within the pollen-chamber, and the pluricellular structure of these pollen-grains has been repeatedly demonstrated. In the light of our present knowledge of Ginkgo and the Cycads, there can scarcely be a doubt that spermatozoids were formed in the cells of the antheridium of the Cordaitean pollen-grain and that of other Palaeozoic Spermophyta; the antheridium is much more developed than in any recent Gymnosperm, and it may be doubted whether any pollen-tube was formed. The morphology of the female inflorescence of Cordaiteae has not yet been cleared up, but Taxus and Ginkgo among recent plants appear to offer the nearest analogies. Much further investigation will be needed before the homologies between Cordaitean cones and the fructifications of the higher Cryptogams can be established. Anatomically the connexion of the family with the Pteridosperms (and through them, presumably, with some primitive group of Ferns) seems clear, but we have as yet no indications of the stages in the evolution of their reproductive organs. The class Cordaitales extends back to the Devonian, and it must be borne in mind that our knowledge of their fructifications is practically limited to representatives from the latest Palaeozoic horizons.

 (All after Renault.)

Fig. 32.—Cordaianthus.

A, C. Williamsoni. Part of longitudinal section of ♀ catkin; 𝑎, axis, showing 𝑣, bundles in tangential section; 𝑏𝑟, bracts; 𝑑, short axillary shoot, bearing a bracteole and a terminal ovule; 𝑖, integument; 𝑛, nucellus of ovule; 𝑜𝑣, another ovule seen from the outside. (✕ about 10.)

B, C. Grand’ Euryi. Nucellus of an ovule; 𝑝.𝑐., pollen-chamber; 𝑠, canal leading to 𝑝.𝑐.; 𝑝, pollen-grains in 𝑝.𝑐.; 𝑝′, do. in canal. (✕ about 30.)

C, C. Grand’ Euryi. Lower part of canal, enlarged; 𝑜, cavity of canal, surrounded by a sheath of cells, dilated towards the bottom of canal, in which a large pollen-grain is caught; 𝑒𝑥, exterior of pollen grain; 𝑖𝑛, internal group of prothallial or antheridial cells. (✕ 150.)

D, Cycadinocarpus augustodunensis. Upper part of seed, in longitudinal section; 𝑖, integument; 𝑚𝑖, micropyle; 𝑛, remains of nucellus; 𝑝.𝑐., pollen-chamber (containing pollen-grains), with its canal extending up to the micropyle; 𝑝𝑟, part of prothallus; 𝑎𝑟, archegonia. All figures magnified.

Isolated fossil seeds are common in the Carboniferous and Permian strata; in all cases they are of the orthotropous type, and resemble the seeds of Cycads or Ginkgo more nearly than those of any other living plants. Their internal structure is sometimes admirably preserved, so that the endosperm with its archegonia is clearly shown (fig. 32, D). It is a curious fact that in no case has an embryo been found in any of these seeds; probably fertilization took place after they were shed, and was followed immediately by germination. There is good evidence that many of the seeds belonged to Cordaitales, especially those seeds which had a flattened form, such as Cardiocarpus, Cycadinocarpus, Samaropsis, &c. Seeds of this kind have been found in connexion with the Cordaianthus inflorescences; the winged seeds of Samaropsis, borne on long pedicels, are attributed by Grand’ Eury to the genus Dorycordaites. Many other forms of seed, and especially those which show radial symmetry, as for example Trigonocarpus, Stephanospermum and Lagenostoma belonged, as we have seen, to some of the plants grouped under Pteridospermeae, though other Pteridosperms had flattened seeds not as yet distinguishable from those of Cordaitales. The abundance and variety of Palaeozoic seeds, still so often of undetermined nature, indicate the vast extent of the spermophytic flora of that period.

The modern Gymnospermous orders have but few authentic representatives in Palaeozoic rocks. The history of the Ginkgoales will be found in the Mesozoic section of this article (see also Gymnosperms); their nearest Palaeozoic representatives “were probably members of the Cordaitales, an extinct stock with which the Ginkgoaceae are closely connected” (Seward). Remains referable to Cycadophyta, so extraordinarily abundant in the succeeding period, are scanty. The curious genus Dolerophyllum (Saporta) may be mentioned in this connexion. This genus, from the Permo-Carboniferous of Autun, is represented by large, fleshy, reniform leaves or leaflets, with radiating dichotomous venation; the vascular bundles have in all respects the structure of those in the leaves of Cycads or Cordaiteae. The male sporophylls are similar in form to the vegetative leaves, but smaller; sunk in their parenchyma are numerous tubular loculi, containing large pollen-grains, which are pluricellular like those of Cordaites; the female fructification had not yet been identified with certainty. The curious male sporophylls may perhaps be remotely comparable to those recently discovered in Mesozoic Cycadophyta, of the group Bennettiteae. Some leaves of Cycadean habit (e.g. Pterophyllum, Sphenozamites) occur in the Coal Measures and Permian, and it is possible that the obscure Coal Measure genus Noeggerathia may have Cycadean affinities. A fructification from the Permian of Autun, named Cycadospadix milleryensis by Renault, appears to belong to this family.

Now that the numerous specimens of wood formerly referred to Coniferae are known to have belonged to distinct orders, but few true Palaeozoic Conifers remain to be considered. The most important are the upper Coal Measure or Permian genera Walchia, Ullmannia and Pagiophyllum, all of which resembled certain Araucarieae in habit. In the case of Walchia there is some evidence as to the fructifications, which in one species (W. filiciformis) appear to be comparable to female Araucarian cones. There are also some anatomical points of agreement with that family. It is probable, however, that under the same generic name very heterogeneous plants have been confounded. In the case of Ullmannia the anatomical structure of the leaf, investigated by Solms-Laubach, proves at any rate that the tree was Coniferous.

There is no proof of the existence of Gnetaceae in Palaeozoic times. The very remarkable plumose seeds described by Renault under the name Gnetopsis are of uncertain affinity, but have much in common with Lagenostoma, the seed of Lyginodendron.

Succession of Floras.

Our knowledge of vegetation older than the Carboniferous is still far too scanty for any satisfactory history of the Palaeozoic Floras to be even attempted; a few, however, of the facts may be advantageously recapitulated in chronological order.

No recognizable plant-remains, if we accept one or two doubtful Algal specimens, have so far been yielded by the Cambrian. From the Ordovician and Silurian, however, a certain number of authentic remains of Algae (among many more that are questionable) have been investigated; they are for the most part either verticillate Siphonae, or the large—possibly Laminariaceous—Algae named Nematophycus, with the problematical but perhaps allied Pachytheca. The evidence for terrestrial Silurian vegetation is still dubious; apart from some obscure North American specimens, the true nature of which is not established, Potonié has described well-characterized Pteridophytes (such as the fern-like Sphenopteridium and Bothrodendron among Lycopods) from supposed Silurian strata in North Germany; the horizon, however, appears to be open to much doubt, and the specimens agree so nearly with some from the Lower Carboniferous as to render their Silurian age difficult of credence. The high development of the terrestrial flora in Devonian times renders it probable that land-plants existed far back in the Silurian ages, or still earlier. Even in the Lower Devonian, Ferns and Lepidodendreae have been recognized; the Middle and Upper Devonian beds contain a flora in which all the chief groups of Carboniferous plants are already represented. Considering the comparative meagreness of the Devonian record, we can scarcely doubt that the vegetation of that period, if adequately known, would prove to have been practically as rich as that of the succeeding age. Among Devonian plants, Equisetales, including not only Archaeocalamites, but forms referred to Asterophyllites and Annularia, occur; Sphenophyllum is known from Devonian strata in North America and Bear Island, and Pseudobornia from the latter; Lycopods are represented by Bothrodendron and Lepidodendron; a typical Lepidostrobus, with structure preserved, has lately been found in the Upper Devonian of Kentucky. Fern-like plants such as Sphenopterideae, Archaeopteris and Aneimites, with occasional arborescent Pecopterideae, are frequent; many of the genera, including Alethopteris, Neuropteris and Megalopteris, probably belonged, not to true Ferns, but to Pteridosperms; although our knowledge of internal structure is still comparatively scanty, there is evidence to prove that such plants were already present, as for example, the genus Calamopitys. The presence of Cordaitean leaves indicates that Gymnosperms of high organization already existed, a striking fact, showing the immense antiquity of this class compared with the angiospermous flowering plants.

Any detailed account of the horizons of Carboniferous plants would carry us much too far. For our present purpose we may divide the formation into Lower Carboniferous and Lower and Upper Coal Measures. In the Lower Carboniferous (Culm of Continental authors) many Devonian types survive—e.g. Archaeocalamites, Bothrodendron, Archaeopteris, Megalopteris, &c. Among fern-like fronds Diplotmema and Rhacopteris are characteristic. Some of the Lepidodendreae appear to approach Sigillariae in external characters. Sphenophylleae are still rare; it is to this horizon that the isolated type Cheirostrobus belongs. Many specimens with structure preserved are known from the Lower Carboniferous, and among them Pteridosperms (Heterangium, Calamopitys, Cladoxylon, Protopitys) are well represented, if we may judge by the anatomical characters. Of Gymnosperms we have Cordaitean leaves, and the stems known as Pitys, which probably belonged to the same family.

The Lower Coal Measures (Westphalian) have an enormously rich flora, embracing most of the types referred to in our systematic description. Calamarieae with the Arthropitys type of stem-structure abound, and Sphenophylleae are now well represented. Bothrodendron still survives, but Lepidodendron, Lepidophloios, and the ribbed Sigillariae are the characteristic Lycopods. The heterogeneous “Ferns” grouped under Sphenopterideae are especially abundant. Ferns of the genera referred to Marattiaceae are common, but arborescent stems of the Psaronius type are still comparatively rare. Numerous fronds such as Alethopteris Neuropteris, Mariopteris, &c., belonged to Pteridosperms, of which specimens showing structure are frequent in certain beds. Cordaites, Dorycordaites and many stems of the Mesoxylon type represent Gymnosperms; the seeds of Pteridosperms and Cordaiteae begin to be common. The Upper Coal Measures (Stephanian) are characterized among the Calamarieae, now more than ever abundant, by the prevalence of the Calamodendreae; new species of Sphenophyllum make their appearance; among the Lycopods, Lepidodendron and its immediate allies diminish, and smooth-barked Sigillariae are the characteristic representatives. “Ferns” and Pteridosperms are even more strongly represented than before, and this is the age in which the supposed Marattiaceous tree-ferns reached their maximum development. Among Pteridosperms it is the family Medulloseae which is especially characteristic. Cordaiteae still increase, and Gymnospermous seeds become extraordinarily abundant. In the Upper Coal Measures the first Cycadophyta and Coniferae make their appearance. The Permian, so far at least as its lower beds are concerned, shows little change from the Stephanian; Conifers of the Walchia type are especially characteristic. The remarkable Permo-Carboniferous flora of India and the southern hemisphere is described in the next section of this article. During the earlier part of the Carboniferous epoch the vegetation of the world appears to have been remarkably uniform; while the deposition of the Coal Measures, however, was in progress, a differentiation of floral regions began. The sketch given above extends, for the later periods, to the vegetation of the northern hemisphere only.

Authorities.—Potonié, Lehrbuch der Pflanzenpaläontologie (Berlin, 1899); Renault, Cours de botanique fossile, vols. i.–iv. (Paris, 1881–1885); Scott, Studies in Fossil Botany (2nd ed., London, 1908–1909); “The present Position of Palaeozoic Botany,” in Progressus rei botanicae, Band I. (Jena, 1907); Seward, Fossil Plants (in course of publication), vol. i. (Cambridge, 1898), vol. ii. (1910); Solms-Laubach, Introduction to Fossil Botany (Oxford, 1892); Zeiller, Éléments de paléobotanique (Paris, 1900). In these general works references to all important memoirs will be found.  (D. H. S.) 

II.—Mesozoic

The period dealt with in this section does not strictly correspond with that which it is customary to include within the limits of the Mesozoic system. The Mesozoic era, as defined in geological textbooks, includes the Triassic, Jurassic and Cretaceous epochs; but from the point of view of the evolution of plants and the succession of floras, this division is not the most natural or most convenient. Our aim is not simply to give a summary of the most striking botanical features of the several floras that have left traces in the sedimentary rocks, but rather to attempt to follow the different phases in the development of the vegetation of the world, as expressed in the contrasts exhibited by a comparison of the vegetation of the Coal period forests with that of the succeeding Mesozoic era up to the close of the Wealden period.

Towards the close of the Palaeozoic era, as represented by the Upper Carboniferous and Permian plant-bearing strata, the vegetation of the northern hemisphere and that of several regions in the southern hemisphere, consisted of numerous types of Vascular Cryptogams, with some members of the Gymnospermae, and several genera referred to the Pteridospermae and Cycadofilices (see section I. Palaeozoic). In the succeeding Permian period the vegetation retained for the most part the same general character; some of the Carboniferous genera died out, and a few new types made their appearance. The Upper Carboniferous and Permian plants may be grouped together as constituting a Permo-Carboniferous flora characterized by an abundance of arborescent Vascular Cryptogams and of an extinct class of plants to which the name Pteridosperms has recently been assigned—plants exhibiting a combination of Cycadean and filicinean characters and distinguished by the production of true gymnospermous seeds of a complex type. This flora had a wide distribution in North America, Europe and parts of Asia; it extended to China and to the Zambesi region of tropical Africa (Map A, I. and II.).

On the other hand, the plant-beds of the Permo-Carboniferous age in South Africa, South America, India and Australia demonstrate the existence of a widely distributed vegetation which agrees in age with the Upper Carboniferous and Permian vegetation of the north, but differs from it to such an extent as to constitute a distinct flora. We must Glossopteris Flora. begin by briefly considering this southern Palaeozoic province if we would trace the Mesozoic floras to their origin, and obtain a connected view of the vegetation of the globe as it existed in late Palaeozoic times and at the beginning of the succeeding era.

Fig. 1.—Glossopteris frond, with portion enlarged to show the venation. (Natural size＝36 cm. in length.) From Lower Gondwana rocks of India.

In Australia, South America and South Africa a few plants have been found which agree closely with Lower Carboniferous types of the northern hemisphere. In New South Wales, for example, we have such genera as Rhacopteris and Lepidodendron represented by species very similar to those recorded from Lower Carboniferous or Culm rocks in Germany, Austria, England, Spitzbergen, North and South America and elsewhere. It is, in short, clear that the Culm flora, as we know it in the northern hemisphere, existed in the extreme south, and it is probable that during the earlier part of the Carboniferous period the vegetation of the world was uniform in character. We may possibly go a step farther, and assume that the climatic conditions under which the Culm plants of the Arctic regions flourished were not very different from those which prevailed in Europe, Asia, Chile and South Australia. From strata in New South Wales overlying Devonian and Lower Carboniferous rocks certain plants were discovered in the early part of the 19th century which were compared with European Jurassic genera, and for several years it was believed that these plant-beds belonged to the Mesozoic period. These supposed Mesozoic plants include certain genera which are of special interest. Foremost among these is the genus Glossopteris (fig. 1), applied by Brongniart in 1828 to sub-lanceolate or tongue-shaped leaves from India and Australia, which have generally been regarded as the fronds of ferns characterized by a central midrib giving off lateral veins which repeatedly anastomose and form a network, like that in the leaves of Antrophyum, an existing member of the Polypodiaceae. The stems, long known from Australia and India as Vertebraria, have in recent years been proved to be the rhizomes of Glossopteris. It is only recently that undoubted sporangia have been found in close association with Glossopteris leaves. The genus possessed small broadly oval or triangular leaves in addition to the large fronds like that shown in fig. 1; it was with the smaller leaves that Mr Arber discovered sporangia exhibiting certain points of resemblance to the microsporangia of modern Cycads. We cannot as yet say whether these bodies represent a somewhat unusual type of fern sporangium or whether they are microsporangia; if the latter supposition is correct the plant must have been heterosporous; but we are still without evidence on this point. Associated with Glossopteris occurs another fern, Gangamopteris, usually recognized by the absence of a well marked midrib, though this character does not always afford a satisfactory distinguishing feature. In view of recent discoveries which have demonstrated the Pteridosperm nature of many supposed ferns of Palaeozoic age, we must admit the possibility that the term fern as applied to Glossopteris and Gangamopteris may be incorrect. An Equisetaceous plant, which Brongniart named Phyllotheca in 1828, is another member of the same flora; this type bears a close resemblance to Equisetum in the long internodes and the whorled leaves encircling the nodes, but differs in the looser leaf-sheaths and in the long spreading filiform leaf-segments, as also in the structure of the cones. Phyllotheca has been recognized in Europe in strata of Palaeozoic age, and Professor Zeiller has discovered a new species—P. Rallii—in Upper Carboniferous rocks in Asia Minor (Map A, VII.), which points to a close agreement between this genus and the well-known Palaeozoic Annularia. Phyllotheca occurs also in Jurassic rocks in Italy and in Siberian strata originally described as Jurassic, but which Zeiller has shown are no doubt of Permian age. Some examples of this genus, described by Etheridge from Permo-Carboniferous beds in New South Wales, differ in some respects from the ordinary form, and bear a superficial resemblance to the Equisetaceous genus Cingularia from the Coal Measures of Germany. Other genera characteristic of this southern flora are mentioned later. The extraordinary abundance of Glossopteris in Permo-Carboniferous rocks of Australia, and in strata of the same age in India and South Africa, gave rise to the term “Glossopteris flora” for the assemblage of plants obtained from southern hemisphere rocks overlying beds containing Devonian and Lower Carboniferous fossils. The Glossopteris flora of Australia occurs in certain regions in association with deposits which are now recognized as true boulder-beds, formed during widespread glacial conditions. In India the same flora occurs in a thick series of fresh-water sediments, known as the Lower Gondwana system, including basal boulder-beds like those of Australia. Similar glacial deposits occur also in South America, and members of the Glossopteris flora have been discovered in Brazil and elsewhere. In South Africa, Glossopteris, Gangamopteris and other genera, identical with those from Australia and India, are abundantly represented, and here again, as in India and South America, the plants are found in association with extensive deposits of undoubted glacial origin. To state the case in a few words: there is in South Africa, South America, Australia and India an extensive series of sediments containing Glossopteris, Gangamopteris and other genera, and including beds full of ice-scratched boulders. These strata are homotaxial with Permo-Carboniferous rocks in Europe and North America, as determined by the order of succession of the rocks, and by the occurrence of typical Palaeozoic shells in associated marine deposits. The most important evidence on which this conclusion is based is afforded by the occurrence of European forms of Carboniferous shells in marine strata in New South Wales, which are intercalated between Coal Measures containing members of the Glossopteris flora, and by the discovery of similar shells, many of which are identical with the Australian species, in strata in the north-west of India and in Afghanistan, forming part of a thick series of marine beds known as the Salt Range group. This group of sediments in the extra-peninsular area of India includes a basal boulder-bed, referred on convincing evidence to the same geological horizon as the glacial deposits of the Indian peninsula (Talchir boulder-beds), South Africa (Eeca boulder-beds), Australia and Tasmania (Bacchus Marsh boulder-beds, &c.), and South America, which are associated with Glossopteris-bearing strata. We have a flora of wide distribution in South Africa, South America, Borneo, Australia, Tasmania and India which is clearly of Permo-Carboniferous age, but which differs in its composition from the flora of the same age in other parts of the world. This flora appears to have abruptly succeeded an older flora in Australia and elsewhere, which was precisely similar to that of Lower Carboniferous age in the northern hemisphere. The frequent occurrence of ice-formed deposits at the base of the beds in which Glossopteris and other genera make their appearance, almost necessitates the conclusion that the change in the character of the vegetation was connected with a lowering of temperature and the prevalence of glacial conditions over a wide area in India and the southern hemisphere. There can be little doubt that the Indian Lower Gondwana rocks, in which the boulder-beds and the Glossopteris flora occur, must be regarded as belonging to a vast continental area of which remnants are preserved in Australia, South Africa and South America. This continental area has been described as “Gondwana Land,” a tract of enormous extent occupying an area, part of which has since given place to a southern ocean, while detached masses persist as portions of more modern continents, which have enabled us to read in their fossil plants and ice-scratched boulders the records of a lost continent in which the Mesozoic vegetation of the northern hemisphere had its birth. Of the rocks of this southern continent those of the Indian Gondwana system are the richest in fossil plants; the most prominent types recorded from these Permo-Carboniferous strata are Glossopteris, Gangamopteris, species referred to Sphenopteris, Pecopteris, Macrotaeniopteris and other Ferns; Schizoneura (fig. 2) and Phyllotheca among the Equisetales, Naeggerathiopsis and Euryphyllum, probably members of the Cordaitales (q.v. in section I. Palaeozoic); Glossozamites and Pterophyllum among the Cycadales, and various vegetative shoots recalling those of the coniferous genus Voltzia, a well-known Permian and Triassic plant of northern latitudes. The genera Lepidodendron, Sigillaria, Stigmaria, or Calamites, which played so great a share in the vegetation of the same age in the northern hemisphere, have not been recognized among the Palaeozoic forms of India, but examples of Sigillaria, Lepidodendron and Bothrodendron are known to have existed in South Africa in the Permo-Carboniferous era.

We may next inquire what types occur in the Glossopteris flora agreeing more or less closely with members of the rich Permo-Carboniferous vegetation of the north. The genus Sphenophyllum, abundant in the Coal Measures and Permian rocks of Europe and America, is represented by a single species recorded from India, Sphenophyllum speciosum (fig. 3), and a doubtful species from South Africa; Annularia, another common northern genus, is recorded from Australia, and the closely allied Phyllotheca constitutes another link between the two Permo-Carboniferous floras. The genus Cordaites may be compared, and indeed is probably identical with, certain forms recorded from India, South America, South Africa and Australia. While a few similar or even identical types may be recognized in both floras, there can be no doubt that, during a considerable period subsequent to that represented by the Lower Carboniferous or Culm rocks, there existed two distinct floras, one of which had its headquarters in the northern hemisphere, while the other flourished in a vast continental area in the south. Recent discoveries have shown that representatives of the two floras coexisted in certain regions; there was, in fact, a dovetailing between the northern and southern botanical provinces. In 1895 Professor Zeiller described several plants from the province of Rio Grande do Sul in South America (Map A, G₂), including a few typical members of the Glossopteris flora associated with a European species, Lepidophloios laricinus, one of the characteristic types of the Coal period, and with certain ferns resembling some species from European Permian rocks. A similar association was found also in Argentine rocks by Kurtz (Map A, G₁), and from South Africa Sigillaria Brardi, Psygmophyllum, Bothrodendron and other northern types are recorded in company with Glossopteris, Glangamopteris and Naeggerathiopsis.

Map A.—G1—G6, Glossopteris Flora.
I, II.	Upper Carboniferous plants of the northern hemisphere facies, in the Zambesi district and in China.
III.	Rhaetic flora of Tongking (Glossopteris, &c.; associated with northern types).
IV.	Carboniferous plants (prov. Kansu).
V.	Glossopteris, &c., in Permian rocks in prov. Vologda.
VI.	Permian (Pechora valley).	VII.	Upper Carboniferous (Herakleion).
VIII.	Rhaetic (Honduras).	IX.	Lower Jurassic, Upper Gondwana (Argentine).
X.	Rhaetic (Persia).	XI.	Triassic—Cretaceous.

---

 (After Feistmantel.)

Fig. 2.—Schizoneura gondwanensis from Lower Gondwana rocks, India.

(After Feistmantel.)
A.	B.
Fig. 3.—Sphenophyllum speciosum. From Lower Gondwana rocks, India.
A. nat. size.B. leaflet enlarged.

The Coal-bearing strata which occupy a considerable area in China (Map A, II.), contain abundant samples of a vegetation which appears to have agreed in their main features with the Permo-Carboniferous floras of the northern hemisphere. In his account of some plants from the Coal Measures of Kansu (Map A, IV.) Dr Krasser has drawn attention to the apparent identity of certain leaf-fragments with those of Naeggerathiopsis Hislopi, a typical member of the Glossopteris flora; but this plant, so far as the evidence of vegetative leaves may be of value, differs in no essential respects from certain species of a European genus Cordaites. A comparatively rich fossil flora was described in 1882 from Tongking (Map A, III. by Professor Zeiller—and this author has recently made important additions to his original account—which demonstrates an admixture of Glossopteris types with others which were recognized as identical with plants characteristic of Rhaetic strata in Europe. In the Tongking area, therefore, a flora existed during the Rhaetic period consisting in part of genera which are abundant in the older Glossopteris beds of the south, and in part of well known constituents of European Rhaetic floras. A characteristic member of the southern botanical province, Schizoneura gondwanensis (fig. 2) of India, is represented also by a closely allied if not an identical species—S. paradoxa—in the Lower Trias (Bunter) sandstones of the Vosges Mountains, associated with European species which do not occur in the Glossopteris flora. Another plant found in the Vosges sandstones—Neuropteridium grandifolium—is also closely allied to species of the same “fern” recorded from the Lower Gondwana strata of India (fig. 4), South America and South Africa. These two instances—the Tongking beds of Rhaetic age and the Bunter sandstones of the Vosges—afford evidence of a northern extension of Glossopteris types and their association with European species. In 1898 an important discovery was made by Professor Amalitzky, which carries us a step further in our search for a connexion between the northern and southern floras. Amalitzky found in beds of Upper Permian age in the province of Vologda (Russia) (Map A, V.) species of Glossopteris and Naeggerathiopsis typical members of the Glossopteris flora, associated with species of the ferns Taeniopteris, Callipteris and Sphenopteris, a striking instance of a commingling in the far north of the northern hemisphere Permian species with migrants from “Gondwana Land.” This association of types clearly points to a penetration of representatives of the Glossopteris flora to the north of Europe towards the close of the Permian period. Evidence of the same northern extension is supplied by floras described by Schmalhausen from Permian rocks in the Pechora valley (Map A, VI.), the Siberian genus Rhiptozamites being very similar to, and probably generically identical with, Naeggerathiopsis of the Glossopteris flora. The Permo-Carboniferous beds of South Africa, India and Australia are succeeded by other plant-bearing strata, containing numerous species agreeing closely with members of the Rhaetic and Jurassic floras of the northern hemisphere. These post-Permian floras, as represented by the Upper Gondwana beds of India and corresponding strata in Australia, South Africa, and South America, differ but slightly from the northern floras, and point to a uniformity in the Rhaetic and Jurassic vegetation which is in contrast to the existence of two botanical provinces during the latter part of the Palaeozoic period. A few plants described by Potonié from German and Portuguese East Africa demonstrate the occurrence of Glossopteris and a few other genera, referred to a Permo-Triassic horizon, in a region slightly to the north of Tete in the Zambesi district (Map A, I.), where typical European plants agreeing with Upper Carboniferous types were discovered several years ago, and described by Zeiller in 1882 and 1901. The existence of Upper Gondwana plants, resembling Jurassic species from the Rajmahal beds of India, has been demonstrated in the Argentine by Dr Kurtz.

(After Feistmantel.)

Fig. 4.—Neuropteridium validum. From Lower Gondwana rocks, India.

Having seen how the Glossopteris flora of the south gradually spread to the north in the Permian period, we may now take a brief survey of the succession of floras in the northern hemisphere, which have left traces in Mesozoic rocks of North America, Europe and Asia. Our knowledge of the Triassic vegetation is far from extensive; this Mesozoic Floras. is no doubt due in part to the fact that the conditions under which the Triassic rocks were deposited were not favourable to the existence of a luxuriant vegetation. Moreover, the Triassic rocks of southern Europe and other regions are typical marine sediments. The Bunter sandstones of the Vosges have afforded several species of Lower Triassic plants; these include the Equisetaceous genus Schizoneura—a member also of the Glossopteris flora—bipinnate fern fronds referred to the genus Anomopteris, another fern, described originally as Neuropteris grandifolia, which agrees very closely with a southern hemisphere type (Neuropteridium validum, fig. 4), some large Equisetaceous stems apparently identical, except in size, with modern Horsetails. With these occur several Conifers, among others Voltzia heterophylla and some twigs referred to the genus Albertia, bearing large leaves like those of Agathis australis and some of the Araucarias, also a few representatives of the Cycadales. Among plants from Lower Triassic strata there are a few which form connecting links with the older Permo-Carboniferous flora; of these we have a species, described by Blanckenhon as Sigillaria oculina, which may be correctly referred to that genus, although an inspection of a plaster-cast of the type-specimen in the Berlin Bergakademie left some doubt as to the sufficiency of the evidence for adopting the generic name Sigillaria. Another Triassic genus, Pleuromeia, is of interest as exhibiting, on the one hand, a striking resemblance to the recent genus Isoetes, from which it differs in its much larger stem, and on the other as agreeing fairly closely with the Palaeozoic genera Lepidodendron and Sigillaria. There is, however, a marked difference, as regards the floras as a whole, between the uppermost Palaeozoic flora of the northern hemisphere and such species as have been recorded from Lower Triassic beds. There is evidence of a distinct break in the succession of the northern floras which is not apparent between the Permian and Trias floras of the south. Passing over the few known species of plants from the middle Trias (Muschelkalk) to the more abundant and more widely spread Upper Triassic species as recorded from Germany, Austria, Switzerland, North America and elsewhere, we find a vegetation characterized chiefly by an abundance of Ferns and Cycads, exhibiting the same general facies as that of the succeeding Rhaetic and Lower Jurassic floras. Among Cycads may be mentioned species of Pterophyllum (e.g. P. Jaegeri), represented by large pinnate fronds not unlike those of existing species of Zamia, some Equisetaceous plants and numerous Ferns which may be referred to such families as Gleicheniaceae, Dipteridinae and Matonineae. Representatives of the Ginkgoales constitute characteristic members of the later Triassic floras, and these, with other types, carry us on without any break in continuity to the Rhaetic floras of Scania, Germany, Asia, Chile, Tonkin and Honduras (Map A, VIII.), and to the Jurassic and Wealden floras of many regions in both the north and south hemispheres. A comparative view of the plants found in various parts of the world, in beds ranging from the Upper Trias to the top of the Jurassic system, reveals a striking uniformity in the vegetation both in northern and southern latitudes during this long succession of ages. The Palaeozoic types are barely represented; the arborescent Vascular Cryptogams have been replaced by Cycads, Ginkgoales and Conifers as the dominant classes, while Ferns continue to hold their own. No undoubted Angiosperms have yet been found below the Cretaceous system. From the close of the Permian period, which marks the limit of the Upper Palaeozoic floras, to the period immediately preceding the apparently sudden appearance of Angiosperms, we have a succession of floras differing from one another in certain minor details, but linked together by the possession of many characters in common. It is impossible to consider in detail this long period in the history of plant-evolution, but we may briefly pass in review the most striking features of the vegetation as exhibited in the dominant types of the various classes of plants. Fragments of a Jurassic flora have recently been discovered by Dr Andersson, a member of Nordenskiold’s Antarctic expedition, in Louis Philippe Land in lat. 63° 15′ S. Among other well-known Jurassic genera Nathorst has identified the following: Equisetites, Cladophlebis, Todites, Thinnfeldia, Otozamites, Williamsonia pecten, Araucarites. The discovery of this Antarctic flora is a further demonstration of the world-wide distribution of a uniform Jurassic flora.

Under the head of Algae there is little of primary importance to record, but it is of interest to notice the occurrence of certain forms which throw light on the antiquity of existing families of Algae. Species referred on good evidence to the Charophyta are represented by a few casts of oögonia and stem fragments, found in Jurassic and Wealden beds, which bear a striking Algae. resemblance to existing species. There is some evidence for the occurrence of similar Chara “fruits” in middle Triassic rocks; some doubtful fossils from the much older Devonian rocks have also been quoted as possible examples of the Charophyta. The oldest known Diatoms are represented by some specimens found entangled in the spicules of a Liassic sponge, and identified by Rothpletz as species of the recent genus Pyxidicula. The calcareous Siphoneae are represented by several forms, identified as species of Diplopora, Triploporella, Neomeris and other genera, from strata ranging from the lower Trias limestones of Tirol to the Cretaceous rocks of Mexico and elsewhere. It is probable that the Jurassic Goniolina, described from French localities, and other genera which need not be mentioned, may also be reckoned among the Mesozoic Siphoneae. A genus Zonatrichites, compared with species of Cyanophyceae, has been described as a Calcareous alga from Liassic limestones of Silesia.

The geological history of Mosses and Liverworts is at present very incomplete, and founded on few and generally unsatisfactory fragments. It is hardly too much to say that no absolutely trustworthy examples of Mosses have so far been found in Mesozoic strata. Of Liverworts there are a few species, such as Palaeohepatica Rostafinskii from the Lower Jurassic Bryophyta. rocks of Cracow, Marchantites erectus from the Inferior Oolite rocks of Yorkshire, and M. Zeilleri from the Wealden beds of Sussex. These fossil Hepaticae are unfortunately founded only on sterile fragments, and placed in the Liverworts on the strength of their resemblance to the thallus of Marchantia and other recent genera.

The Palaeozoic Calamites were succeeded in the Triass'ic period by large Equisetites, differing, so far as we know, in no essential respect from existing Equisetums. The large stems represented by casts of Triassic age, Equisetites arenaceus other species, probably possessed the power of secondary growth in thickness; the cones were of the modern type, Equisetaceae. and the rhizomes occasionally formed large underground tubers like those frequently met with in Equisetum arvense, E. sylvaticum and other species. Equisetites Muensteri is a characteristic and fairly widely spread Rhaetic and Liassic species, having a comparatively slender stem, with leaf-sheaths consisting of a few broad and short leaf-segments. Equisetites columnaris, a common fossil in the Jurassic plant-beds of the Yorkshire coast, represents another type with relatively stout and occasionally branched vegetative shoots, bearing leaf-sheaths very like those of Equisetum maximum and other Horsetails. In the Wealden strata more slender forms have been found—e.g. Equisetites Burchardti and E. Lyelli—in England, Germany, Portugal, Japan and elsewhere, differing still less in dimensions from modern species. Of other Equisetales there are Schizoneura and Phyllotheca; the former first appears in Lower Gondwana rocks as a member of the Glossopteris flora, migrating at a later epoch into Europe, where it is represented by a Triassic species. The latter genus ranges from Upper Carboniferous to Jurassic rocks; it occurs in India, Australia, and elsewhere in the “Gondwana Land” vegetation, as well as in Palaeozoic rocks of Asia Minor, in Permian rocks of Siberia, and in Jurassic plant-beds of Italy. This genus, like the allied Calamites, appears to have possessed cones of more than one type; but we know little of the structure of these Mesozoic Equisetaceous genera as compared with our much more complete knowledge of Calamites and Archaeocalamites. (See section I., Palaeozoic.)

Reference has already been made to Sigillaria oculina and to the genus Pleuromeia. Palaeobotanical literature contains several records of species of Lycopodites and Selaginellites; nearly all of them are sterile fragments, bearing a more or less close resemblance to living Club-Mosses and Selaginellas, but lacking the more important reproductive organs. Lycopodiates. Nathorst has recently described a new type of lycopodiaceous cone, Lycostrobus Scotti, from Rhaetic rocks of Scania, from which he obtained both megaspores and microspores. An investigation by Miss Sollas of a plant long known from Rhaetic rocks in the Severn valley as Naiadita acuminata has shown that this genus is in all probability a small lycopodiaceous plant, and neither a Moss nor a Monocotyledon, as some writers have supposed. One of the best-known European species is Lycopodites falcatus, originally described by Lindley and Hutton from the Inferior Oolite of Yorkshire.

Fig. 5.

A, Otozamites Beani. B, O. Bunburyanus. Inferior Oolite, England.

Among the large number of Mesozoic Ferns there are several species founded on sterile fronds which possess but little interest from a botanical standpoint. Some plants, again, have been referred by certain authors to Ferns, while others have relegated them to the Cycads. As examples of these doubtful forms may be mentioned Thinnfeldia, characteristic of Rhaetic and Lower Filicales. Jurassic rocks; Dichopteris, represented by some exceptionally fine Jurassic specimens, described by Zigno, from Italy; and Ctenis, a genus chiefly from Jurassic beds, founded on pinnate fronds like those of Zamia and other Cycads, with linear pinnae characterized by anastomosing veins. Plants referred to Schimper’s genus Lomatopteris and to Cycadopteris of Zigno afford instances of the difficulty of distinguishing between the foliage of Ferns and Cycads. The close resemblance between specimens from Jurassic rocks placed in one or other of the genera Thinnfeldia, Dichopteris, Cycadopteris, &c., illustrates the unsatisfactory custom of founding new names on imperfect fronds. It is of interest to note that some leaf-fragments recently found in Permian rocks of Kansas, and placed in a new genus Glenopteris, are hardly distinguishable from specimens of Jurassic and Rhaetic age referred to Thinnfeldia and other Mesozoic genera. The difficulty of distinguishing between Ferns and Cycads is a necessary consequence of the common origin of these two classes; in Palaeozoic times the Cycadofilicies and Pteridospermae (see section I., Palaeozoic) played a prominent part, and even among recent Cycads and Ferns we still see a few indications of their close relationship. There is reason to believe that compound or generalized types—partly Ferns and partly Cycads—persisted into the Mesozoic era; but without more anatomical knowledge than we at present possess, it is impossible to do more than to point to a few indications afforded by external, and to a slight extent by internal structure, of the survival of Cycadofilicinean types. The genus Otozamites, which it is customary and probably correct to include in the Cycadales, is represented by certain species, such as Otozamites Beani (fig. 5, A), a characteristic Yorkshire fossil of Jurassic age, which in the form of the frond, bearing broad and relatively short pinnae, exhibits a striking agreement with the sterile portions of the fronds of Aneimia rotundifolia, a member of the fern family Schizaeaceae. Again, another species of the same genus, O. Bunburyanus (fig. 5, B), suggests a comparison with fern fronds like that of the recent species Nephrolepis Duffi. The scaly ramenta which occur in abundance on the leaf-stalk bases of fossil Cycads constitute another fern-character surviving in Mesozoic Cycadales. Without a fuller knowledge of internal structure and of the reproductive organs, we are compelled to speak of some of the Mesozoic plants as possibly Ferns or possibly Cycads, and not referable with certainty to one or other class. It has been found useful in some cases to examine microscopically the thin film of coal that often covers the pinnae of fossil fronds, in order to determine the form of the epidermal cells which may be preserved in the carbonized cuticle; rectilinear epidermal cell-walls are usually considered characteristic of Cycads, while cells with undulating walls are more likely to belong to Ferns. This distinction does not, however, afford a safe guide; the epidermal cells of some ferns, e.g. Angiopteris, have straight walls, and occasionally the surface cells of a Cycadean leaf-segment exhibit a fern-like character. Leaving out of account the numerous sterile fronds which cannot be certainly referred to particular families of Ferns, there are several genera which bear evidence in their sori, and to some extent in the form of the leaf, of their relationship to existing types.

The abundance of Palaeozoic plants with sporangia and sori of the Marattiaceous type is in striking contrast to the scarcity of Mesozoic ferns which can be reasonably included in the Marattiaceae. One of the few forms so far recorded is that known as Marattia Muensteri from Rhaetic localities in Europe and Asia. Some species included in the genus Marattiaceae. Danaeites or Danaeopsis from Jurassic rocks of Poland, Austria and Switzerland may possibly be closely allied to the recent tropical genus Danaea. Of the Ophioglossaceae there are no satisfactory examples; one of the few fossils compared with a recent species, Ophioglossum palmatum, was described several years ago from Triassic rocks under the name Cheiropteris, but the resemblance is one of external form only, and practically valueless as a taxonomic criterion. It would appear that the eusporangiate Ferns suddenly sank to very subordinate position after the Palaeozoic era.

Fig. 6.—Cladophlebis denticulata. Inferior Oolite, England.

The Osmundaceae, represented by a few forms of Palaeozoic age, played a more prominent part in the Mesozoic floras. A species described by Schenk from Rhaetic rocks of Franconia as Acrostichites princeps is hardly distinguishable from Todites Williamsoni, a widely distributed species in Inferior Oolite strata. This Jurassic species bore bipinnate fronds Osmundaceae. not unlike those of the South African, Australian, and New Zealand Fern Todea barbara, which were characterized by a stout rachis and short broad pinnules bearing numerous large sporangia covering the under surface of the lamina. Specimens of Todites have been obtained from England, Poland, and elsewhere, sufficiently well preserved to afford good evidence of a correspondence in the structure of their sporangia with those of recent Osmundaceae. This Jurassic and Rhaetic type occurs in England, Germany, Poland, Italy, East Greenland, North America, Japan, China and Persia (Map A, X.). Bipinnate sterile fronds of Todites have in some instances been described under the designation Pecopteris whitbiensis. This and other names, such as Asplenium whitbiense, A. nebbense, Asplenites Roesserti, &c., have been given to bipinnate fronds of a type frequently met with in different genera and families of recent Ferns, e.g. Onoclea Struthiopteris, species of Cyathea, Asplenium, Gymnogramme, &c. In most cases the Rhaetic, Jurassic and Wealden Ferns included under one or other of these names are sterile, and cannot be assigned to a particular family, but some are undoubtedly the leaves of Todites, a genus which may often be recognized by the broad and relatively short bluntly-terminated pinnules. The Jurassic species Cladophlebis denticulata (fig. 6), recorded from several European localities, as well as from North America, Japan, China, Australia, India and Persia, affords an instance of a common type of bipinnate frond similar to Todites Williamsoni, which has been included in the Polypodiaceae; but such meagre evidence of the soral characters as we possess also points to a comparison with the recent fern Todea barbara. Our knowledge of the anatomy of fossil Osmundaceae has recently been considerably extended by Kidston and Gwynne-Vaughan. (For references, see Seward, Fossil Plants, vol. ii., 1910.)

Fig. 7.—Klukia exilis.

1–3, Sporangia enlarged.

4, Single fertile pinnule slightly
 enlarged.

5, Fragment of pinna.
   Inferior Oolite, England.

The Schizaeaceae include a widely spread species, originally named Pecopteris exilis, and subsequently placed in a new genus, Klukia (fig. 7), which is characterized by tripinnate fronds with short linear ultimate segments, bearing a single row of sporangia with an apical annulus (“monangic sori” of Prantl) on either side of the midrib. This type occurs in Rhaetic Schizaeaceae. and Lower Jurassic rocks of England, the Arctic regions, Japan and elsewhere. Ruffordia Goepperti, a Wealden type, and probably a member of the Schizaeaceae, has been recorded from England, Belgium, and other European countries, and Japan.

The Gleicheniaceae appear to have been represented by Triassic species in North America and Europe, and more abundantly in Jurassic, Wealden, or Lower Cretaceous rocks in Belgium, Greenland, Poland and elsewhere. Some exceptionally perfect fragments of rhizomes have been found by Dr C. Bommer of Brussels in some Wealden deposits at Hainaut in Belgium; but these have not yet been fully described. The dichotomously-branched fronds of the type represented by several recent species of Gleichenia, e.g. Gleicheniaceae. G. dichotoma, &c., are abundant in Lower Cretaceous plant-beds of Greenland, and suggest that in the latter part of the Mesozoic period the Gleicheniaceae held a position in the vegetation of the far north similar to that which they now occupy in the southern tropics of India and other regions.

Fig. 8.—Matonidium Goepperti.

A. Summit of petiole.

B, Fertile pinnules.

Inferior Oolite, England.

The recent Malayan genus Matonia (Map B, Matonia), represented by two species, M. pectinata and M. sarmentosa, is clearly a survival in southern latitudes of a family which occupied an important place in the vegetation of the Rhaetic, Jurassic and Wealden periods. The genera Lacopteris and Matonidium Matonineae. (fig. 8) may be cited as the two most important types, both as regards geographical and geological range, of this Mesozoic family; these ferns are recorded from England, France, Belgium, Germany, Austria, Portugal, Poland and Italy (Map B, M₁), also from Greenland (Map B, M₂), Spitsbergen (Map B, M₃), and Persia (Map B, M₄). From the southern Hemisphere, on the other hand, we know of one or two fragments only which can reasonably be referred to the Matonineae (Map B, M₅), a fact which may point to a northern origin for this family with its two surviving species almost confined to the Malayan region.

The recent genus, Dipteris, with its four existing species, occurring chiefly in the Indo-Malayan region (Map B, Dipteris), is also a modern survival of several Mesozoic types represented by such genera as Dictyophyllum (fig. 9), Hausmannia and Camptopteris, which were abundant during the Rhaetic and Jurassic periods in England, Germany, Sweden and elsewhere in Europe (Map B, D). Important additions to our Dipteridinae. knowledge of the fertile leaves and rhizomes of certain Rhaetic species of Dictyophyllum and other genera have recently been made by Professor Nathorst of Stockholm, and Professor Richter of Quedlinburg has made a thorough investigation of the vegetative organs of Hausmannia, a genus possibly identical with Protorhipis, which is abundant in Lower Cretaceous and other strata in various European localities. The Dipteridinae are represented also by species from Mesozoic rocks of Persia (Map B, D₂), Greenland (Map B, D₃), North America (D₄), South America (D₅) and China (D₆).

(After Schenk.)

Fig. 9.—Dictyophyllum. Rhaetic rocks of Europe and Asia.

The Cyatheaceae constitute another family of leptosporangiate Ferns which had several representatives in Mesozoic floras. The numerous species of fronds from Jurassic and Wealden rocks of North America and Europe referred to Thyrsopteris, a recent monotypic genus confined to Juan Fernandez, are in Cyatheaceae. the majority of cases founded on sterile leaves, and of little or no botanical value. On the other hand, there are several fossil Ferns of Jurassic age possessing cup-like sori like those of Thyrsopteris and other Cyatheaceous Ferns, which indicate a wide Mesozoic distribution for this family. Among Jurassic species which should probably be classed as Cyatheaceae, Coniopteris hymenophylloides is recorded from England, France, Russia, Poland, Bornholm, Italy, the Arctic regions, North America, Japan, China, Australia and India. A few tree-ferns which may be included in this family—such as Protopteris—have been described from Wealden and Lower Cretaceous rocks of England, Germany and Austria. It is by no means easy in dealing with fossil ferns to distinguish between certain Polypodiaceae—such as species of Davallia—and members of the Cyatheaceae.

It is a striking fact that among the numerous Mesozoic Ferns there are comparatively few that can with good reason be referred to the Polypodiaceae, a family which plays so dominant a rôle at the present day. The frequent occurrence of such names as Asplenium, Adiantum, Davallia, and other Polypodiaceous genera in lists of fossil ferns is thoroughly Polypodiaceae. misleading. There are, indeed, a certain number of species which show traces of sori like those of modern species of Asplenium and other genera, but in most cases the names of recent ferns have been used on insufficient grounds. The Wealden and Jurassic genus, Onychiopsis of England, Portugal, Belgium, Germany, Japan, South Africa and Australia, bears a close resemblance to the recent Onychium (Cryptogamme). Other Jurassic Ferns described by Raciborski from Poland suggest a comparison with Davallia. The resemblance of the sporocarp-like bodies—discovered by Nathorst in association with Rhaetic Sagenopteris leaves, and more recently figured by Halle under a new generic name (Hydropterangium)—to the sporocarps of Marsilia is an argument in favour of including Sagenopteris in the Hydropterideae. The majority of the specimens included in the genus Cladophlebis, the Mesozoic representative of the Palaeozoic Pecopteris type of frond, are known only in a sterile condition, and cannot be assigned to their family position. A Wealden plant, Weichselia Mantelli, is worthy of mention as a species of very wide geographical distribution, and one of the most characteristic members of the Wealden flora. This type is distinguished by its large bipinnate fronds bearing long and narrow pinnae with close-set pinnules, characterized by the anastomosing secondary veins. No traces of sori have so far been found on the fronds. Similarly, the genus Sagenopteris, characterized by a habit like that of Marsilia, and represented by fronds consisting of a few spreading broadly oval or narrow segments, with anastomosing veins, borne on the apex of a common petiole, is abundant in rocks ranging from the Rhaetic to the Wealden, but has not so far been satisfactorily placed. The evidence adduced by Nathorst and some other writers is, however, not convincing; until we find well-preserved sporocarps in connection with vegetative fronds we prefer to keep an open mind as regards the position of Sagenopteris.

The abundance of Cycadean plants is one of the most striking features of Mesozoic floras. In most cases we have only the evidence of sterile fronds, and this is necessarily unsatisfactory; but the occurrence of numerous stems and fertile shoots demonstrates the wealth of Cycadean plants in many parts of the world, more particularly during the Jurassic and Wealden Cycadales. periods. From Palaeozoic rocks a few fronds have been described, such as Pterophyllum Fayoli, P. Combrayi, Plagiozamites and Sphenozamites, chiefly from French localities, which are referred to the Cycads because of their similarity to the pinnate fronds of modern Cycadaceae. In the succeeding Triassic system Cycadean plants become much more abundant, especially in the Keuper period; from Rhaetic rocks a still greater number of types have been recorded, among which may be mentioned Nilssonia (fig. 10), Anomozamites, Pterophyllum, Otozamites, Cycadites (fig. 11). The species of Nilssonia shown in fig. 10 (N. compta) is a characteristic member of the Jurassic flora, practically identical with a form from Rhaetic rocks described as Nilssonia polymorpha. The large frond of Cycadites represented in fig. 11 (C. Saportae) is from the Wealden strata of Sussex, and possibly identical with Cycadites tenuisectus from Portugal. In addition to these genera there are others, such as Ctenozamites, Ctenis, and Podozamites, the position of which is less certain. Ctenozamites occurs chiefly in the Rhaetic coal-bearing beds of Scania, and has been found also in the Liassic clays of Dorsetshire and in the Inferior Oolite beds of Yorkshire, as well as in Rhaetic strata in Persia and elsewhere; it is characterized by its bipinnate fronds, and may be compared with the recent Australian genus Bowenia—peculiar among living Cycads in having bipinnate fronds. Ctenis has been incorrectly placed among the ferns by some authors, on account of the occurrence of supposed sporangia on its pinnae; but there is reason to believe that these so-called sporangia are probably nothing more than prominent papillose cells of the epidermis. Podozamites (fig. 12) is usually considered to be a Cycad, but the broad pinnae (or leaves) and their arrangement on the axis suggests a possible relationship with the southern coniferous genus Agathis, represented by the Kauri pine and other recent species. The considerable variation in the size of the pinnae of Podozamites, as represented by species from the Jurassic rocks in the Arctic regions and various European localities, recalls the variation in length and breadth of the leaves of Agathis. With regard to the distinguishing features and the distribution of the numerous Cycadean leaves of Mesozoic age, the most striking fact is the abundance of fronds, which there is good reason to refer to the Cycadales in Upper Triassic, Rhaetic, Jurassic and Wealden rocks in India, Australia, Japan, China and elsewhere in the southern hemisphere, as well in North America, Greenland, and other Arctic lands and throughout Europe. It is noteworthy that Tertiary plant-beds have yielded hardly any specimens that can be recognized as Cycads.

Map B.—M₁–M₅, D, G, Distribution of the Matonineae, Dipteridinae, Ginkgoales.

D1–D6, Distribution of the Dipteridinae.	G1–G17, Distribution of the Ginkgoales during the Mesozoic and Tertiary Periods.
G1 (Trias-Tertiary);	G2, G3 (Rhaetic-Jurassic);
G4 (Tertiary, Sakhalin I.);	G5 (Jurassic);
G6 (Jurassic and Tertiary);	G7 (Jurassic);
G8 (Rhaetic-Jurassic);	G9 (Trias-Rhactic);
G10 (Rhaetic, Chile);	G11 (Trias);
G12 (Cretaceous-Tertiary);	G13 (Tertiary, Alaska);
G14 (Cretaceous-Tertiary);	G15 (Jurassic);
G16 (Jurassic, Spitsbergen);	G17 (Jurassic, Franz Josef Land).

Fig. 10.—Nilssonia compta. Inferior Oolite, England.

Fig. 11.—Cycadites Saportae.
 Wealden, England.

A more important question is, What knowledge have we of the reproductive organs and stems of these fossil Cycads? Cycadean stems have recently been found in great abundance in Jurassic and possibly higher strata in Wyoming, South Dakota and other parts of the United States. Cycadean stems have been found also in the uppermost Jurassic, Wealden and Lower Cretaceous rocks of England, India and other parts of the world. An example of an Indian Cycadean stem from Upper Gondwana rocks is represented in fig. 13; the surface of the trunk is covered with persistent bases (fig. 13, A) of the fronds known as Ptilophyllum cutchense, which are practically the same as the European species Williamsonia pecten (fig. 17). In a section of the stem (fig. 13, B) a large pith is seen to occupy the axial region, and this is surrounded by a zone of secondary wood, which appears to differ from the characteristic wood of modern Cycads (see Gymnosperms) in having a more compact structure. It is interesting to find that G. R. Wieland of Yale University has noticed in some of the Cycadean stems from the Black hills of Dakota and Wyoming that the wood appears to possess a similar structure, differing in its narrower medullary rays from the wood of modern Cycads. The lozenge-shaped areas external to the axis of the stem represent the sections of petioles, some of which are shown in fig. 13, A, attached to the stem. The majority of Mesozoic stems agree in external appearance with those of recent species of Encephalartos, Macrozamia, and some other genera; the trunk is encased in a mass of persistent petiole-bases separated from one another by a dense felt or packing of scaly ramenta. The structure of the leaf-stalks is like that of modern Cycads, but the ramenta, instead of having the form of long unicellular hairs like those on the petioles and bud-scales of existing species are exactly like the paleae or ramental scales characteristic of the majority of ferns. This fern-like character affords an interesting survival of the close relationship between Cycads and Ferns. Some examples of Jurassic Cycadean stems from Wyoming are characterized by an unusually rich development of ramental scales; the ramenta from the old leaf-bases form an almost complete covering over the surface of the trunk. Professor Lester Ward has instituted a new generic name, Cycadella, for these woolly forms. In a few cases the fossil stems show no trace of any lateral flowering shoots, and in that respect agree with modern forms: an instance of this is afforded by a large Cycadean trunk discovered some years ago in one of the Portland quarries, and named Cycadeoidea gigantea (fig. 14). In this stem the flowers may have been terminal, as in existing Cycads. As a rule, however, the fossil stems show a marked difference from modern forms in the possession of lateral shoots given off from the axils of leaves, and terminating in a flower of complex structure containing numerous orthotropous seeds. These reproductive shoots differ in many important respects from the flowers of recent Cycads, and chiefly on this account it is customary to include the plants in a separate genus, Bennettites, and in a separate group—the Bennettitales—distinct from that of the Cycadales including the existing Cycads. The best preserved specimens of the true Bennettites type so far described are from the Lower Greensand and Wealden of England, and from Upper Mesozoic strata in North America, Italy and France. A study of the anatomical structure of the vegetative stem, which on the whole is very similar to that of recent Cycads (fig. 15, 1 and 2), reveals certain characters which are not met with in modern Cycads. The chief distinguishing feature afforded by the leaf-traces; in recent species (see Gymnosperms) these pursue a somewhat complicated course as they pass from the petiole towards the vascular cylinder of the stem, but in Bennettites the vascular bundles from the leaves followed a more direct course through the cortex of the stem (fig. 15, 3). Among existing types the genus Macrozamia appears to show the nearest approach to this simpler structure of the leaf-traces. In a Floridan species of Zamia the leaf-traces are described as characterized by a more direct course from the stele of the stem to the leaves than in most modern genera, thus agreeing more closely with the extinct Bennettites. The typical Bennettites female flower (fig. 15, 4 and 7), as investigated in English, French, Italian, and American specimens, may be briefly described as a short lateral shoot or peduncle, arising in a leaf-axil and terminating in a bluntly rounded apex, bearing numerous linear bracts enclosing a central group of appendages, some of which consist of slender pedicels traversed by a vascular strand and bearing a single terminal ovule enclosed in an integument, which forms a distal canal or micropyle. Associated with these seminiferous pedicels occur sterile appendages consisting of slender stalks, terminating in distal expansions, which form a fleshy covering over the surface of the flower, leaving small apertures immediately above the micropyles for the entrance of the pollen-grains. It has been suggested by some authors that the almost complete investment of the small Bennettites seeds by the surrounding swollen ends of the interseminal scales (fig. 15, 7) represents an approach to the angiospermous ovary. In Bennettites the ovules are left exposed at the apex, but they are by no means so distinctly gymnospermous as in recent Cycads and Conifers. The seeds have in some cases been preserved in wonderful perfection, enabling one to make out the structure of the embryo, with its bluntly conical radicle and two fleshy cotyledons filling the exalbuminous seed (fig. 15, 11).

Fig. 12.—Podozamites lanceolatus. Inferior Oolite, England.		Fig. 13.—Cycadean stem, from Upper Gondwana rocks, India. A, Surface view; B, Transverse section of stem.		Fig. 14.—Cycadeoidea gigantea. Portland rocks, England.

Our knowledge of the reproductive organs of the Bennettitaceae has until recently been confined to the female flowers, as described by Carruthers, Solms-Laubach, Lignier, and others. The fortunate discovery of several hundred Cycadean stems in the United States, of Lower Cretaceous and Upper Jurassic age, has supplied abundant material which has lately been investigated and is still receiving attention at the hands of Mr Wieland. This investigator has already published a well-illustrated account of his discoveries, which give valuable information as to the morphology of the male organs, and lead us to expect additional results in the future of the greatest importance and interest. On some of the American stems flowers have been found, borne at the apex of lateral shoots, which possess fully developed male organs consisting of sporangia with spores (pollen-grains), surrounding a conical central receptacle bearing numerous small and probably functionless or immature ovules (fig. 15, 10). The structure of this type of flower may be briefly described as follows. In shape and size the flower is similar to that long known as the female flower of Bennettites and Williamsonia. A number of hairy linear bracts enclose the whole; internal to these occur 12 to 20 crowded pinnate leaves (sporophylls), with their apical portions bent over towards the axis of the flower, the bases of the petioles being fused laterally into a disk surrounding the base of the conical receptacle. Numerous pairs of pinnules are attached to the rachis of each sporophyll, and the larger pinnules bear 20 to 30 synangia (sori or plurilocular sporangia) (fig. 15, 8 and 9). The synangia consist of a stout wall composed of thick-walled cells, succeeded by a layer of more delicate and smaller elements; and internal to the wall occur two rows of sporangial loculi containing microspores. When the synangia are ripe dehiscence takes place along a median line between the two rows of loculi. In size, position, arrangement, and manner of dehiscence the sporangia bear a striking resemblance to those of Marattia and Danaea among recent Marattiaceae. The most important point elucidated by this discovery is the very close correspondence of the male organs of the Bennettites flower with the sporophylls and synangia of Marattiaceous ferns—a further relic of the common origin of Cycads and Ferns. It remains to be seen if the ovuliferous cone in the centre of the flower represents simply a functionless gynoecium, as in Welwitschia and abnormal cones of certain Coniferae, or if the flowers were hermaphrodite, with both male and female organs fully developed. We have a combination in the same flower of stalked ovules, the structure of which has already been described, and interseminal scales constituting a complex gynoecium, which exhibits in certain features an approach to the angiospermous type, and differs in structure from other Gymnosperm flowers, associated with male organs constructed on a plan almost identical with that of the sporophylls in Marattiaceae. In many of the flowers described by Mr Wieland the structure is identical in essential features with that of the female flowers of Bennettites Gibsonianus described by Carruthers and by Solms-Laubach, and with that of a French Liassic species described by Lignier: the whole consists of a convex receptacle bearing mature seeds at the tips of pedicels associated with interseminal scales (fig. 15, 7) as already described. Mr Wieland’s researches have, however, demonstrated the existence in flowers of this type of the remains of a disk at the base of the receptacle, between the receptacle and the surrounding bracts, to which staminate leaves were originally attached. The flowers hitherto regarded as female were in some cases at least hermaphrodite, but the male organs had been thrown off before the complete development of the gynoecium. This fact suggests the possibility that the flowers described by Mr Wieland, in which the male organs are mature and the gynoecium is composed of very short and immature ovuliferous stalks and interseminal scales, are not essentially distinct from those which have lost the staminate leaves and possess mature seeds. It is probable that the flowers of Bennettites were normally hermaphrodite, and they may have been markedly protandrous. We cannot decide at present whether the gynoecium in a flower, such as that represented in fig. 15, 7, has partially aborted or whether it would have matured later after the fall of the male organs.

Fig. 15.

1, Bennettites stem: portion of transverse section of stem; 𝑎, vascular cylinder; 𝑏, leaf-traces; 𝑐, pith; 𝑑, cortex.

2, Bennettites stem, tangential section; 𝑒, flower-peduncles.

3, Bennettites stem, leaf-traces attached to the vascular cylinder and passing as simple strands through the cortex;
  𝑑, cortex.

4, Williamsonia, Wealden, England.

5, Young leaf of Bennettites.

6, Ramenta of Bennettites in transverse section.

7, Bennettites, female flower in longitudinal section; 𝑓, apex of peduncle; 𝑔, bracts (shown in surface view in 4);
  ℎ, seeds and seminiferous pedicels; 𝑖, interseminal scales.

8, Bennettites, synangium of male flower, showing line of dehiscence, 𝑘, and microspores, 𝑙.

9, Synangium, in transverse section, showing sporangial groups, 𝑚, and microspores, 𝑙.

10, Bennettites flower in vertical section, showing the central female portion, 𝑛, two sporophylls bearing synangia
  (male), 𝑜, and hairy bracts, 𝑔.

11, Bennettites seed in longitudinal section, showing the dicotyledonous embryo; 𝑝, cotyledons; 𝑟, radicle; 𝑠, testa.

(1–3, after Carruthers; 5, 8, 9 and 10, after Wieland; 7, after Scott; 11, after Solms-Laubach.)

Fig. 16.—Frond of Williamsonia gigas.
Inferior Oolite, England.

It is clear that Bennettites differed in many essential respects from the few modern survivors of the Cycadophyta. Fossil flowers of a type more like that of modern Cycads are few in number, and it is not by any means certain that all of those described as Cycadean flowers and seeds were borne by plants which should be included in the Cycadophyta; a few female flowers have been described from Rhaetic rocks of Scania and elsewhere under the name Zamiostrobus—these consist of an axis with slender pedicels or carpophylls given off at a wide angle and bearing two ovules at the distal end; the structure is in fact similar to that of a Zamia female flower, in which the internodes of the peduncle have been elongated so as to give a looser arrangement to the carpels. It has been suggested that one at least of the flowers, that originally described by Mr Carruthers from the Inferior Oolite of Yorkshire as Beania gracilis, may have been borne by a member of the Ginkgoales. From Jurassic rocks of France and Italy a few imperfect specimens have been described as carpels of Cycads, like those of the recent genus Cycas (see Gymnosperms); while a few of these may have been correctly identified, an inspection of some of the original examples in the Paris collections leads one to express the opinion that others are too imperfect to determine. Pinnate fronds of the Cycas type, characterized by the presence of a midrib and no lateral veins in the linear pinnae, are recorded from Rhaetic rocks of Germany, from Wealden strata in England (fig. 11) and Portugal, and from Liassic beds in Dorsetshire. One large specimen is figured by Heer from Lower Cretaceous rocks of Greenland, and by the side of the frond is shown a carpel with lateral ovules, as in the female flower of Cycas; but an examination of the type-specimen in the Copenhagen Museum led the present writer to regard this supposed carpel as valueless. Professor Nathorst, as the result of a more recent examination of Heer's specimen, found that the segments of the frond are characterized by the presence of two parallel veins instead of a single midrib, with a row of stomata between them; for this type of Cycadean leaf he proposed the generic name Pseudocycas. Another well-known Cycadean genus is Williamsonia, so named by Mr Carruthers in 1870, and now applied to certain pinnate fronds—e.g. those previously described as Zamites gigas (fig. 16), and others known under such names as Pterophyllum or Ptilophyllum pecten, &c., both common Jurassic species—as well as to stems bearing peduncles with terminal oval flowers, similar in form to those of Bennettites. There is good evidence for supporting Professor Williamson's conclusions as to the organic connexion between the flowers, originally described from Inferior Oolite rocks of Yorkshire and subsequently named Williamsonia (fig. 15, 4), and the fronds of Zamites gigas, now known as Williamsonia gigas (fig. 16). There can be little doubt that the majority of the Cycadean fronds of Jurassic and Wealden age, which are nearly always found detached from the rest of the plant, were borne on stems of the Bennettites type. Williamson was the first to express the opinion that the Bennettitean flowers known as Williamsonia were borne on the trunks which terminated in a crown of pinnate fronds of the type long known as Zamites gigas; this view was regarded by Saporta and others as incorrect, and the nature of the Bennettitean foliage was left an open question. A re-examination of the English material in the museums of Paris and elsewhere has confirmed Williamson's conclusions. Mr Wieland has also described young bipinnate fronds, very like those of recent species of Zamia and Encephalartos, attached to a Bennettites stem, and exhibiting the vernation characters of many recent Cycads (fig. 15, 5). In Williamsonia the stem bore comparatively long fertile shoots, which, in contrast to those of Bennettites, projected several inches beyond the surface of the main trunk, and terminated in a flower which appears to have resembled those of the true Bennettites. Nathorst has recently described specimens of Williamsonia from the Jurassic rocks of Whitby with micro-Sporophylls like those of Wieland's species. Williamsonia occurs in the Upper Gondwana rocks of India; it is recorded also from strata ranging from the Rhaetic to the Lower Cretaceous period in England, Portugal, Sweden, Bornholm, Greenland, Italy and North America. Professor Nathorst has described another type of stem from the Rhaetic beds of Scania. It consists of a comparatively small and repeatedly forked axis bearing in each fork a flower; the flowers, which are regarded as male and female, appear to be similar to those of Bennettites. The leaves, borne on the regions between the false dichotomies, are those of Anomozamites minor, a type of Cycadean frond originally determined by Brongniart. The flowers, or some of them, were originally described by Nathorst as Williamsonia angustifolia. This form of stem, of a habit entirely different from that of recent Cycads and extinct Bennettites, points to the existence in the Mesozoic era of another type of Gymnosperm allied to the Bennettitales of the Jurassic and Cretaceous periods by its flowers, but possessing a distinctive character in its vegetative organs. There is no doubt that the Cycadophyta, using the term suggested by Nathorst in 1902, was represented in the Mesozoic period by several distinct families or classes which played a dominant part in the floras of the world before the advent of the Angiosperms. In addition to the bisporangiate reproductive shoots of Bennettites, distinguished by many important features from the flowers of recent Cycads, a few specimens of flowers have been discovered exhibiting a much closer resemblance to those of existing Cycads, e.g. Androstrobus Balduini from Bathonian rocks of France; Zamites familiaris, described many years ago by Corda, from Lower Cretaceous rocks of Bohemia, and Androstrobus Nathorsti, from Wealden beds in Sussex. The majority of the species were, however, characterized by flowers of a different type known as Bennettites and Williamsonia.

Fig. 17.—Fronds of Williamsonia pecten.

The living Maidenhair-tree (Ginkgo biloba) (see Gymnosperms) remains, like Matonia and Dipteris, among the ferns, as an isolated relic in the midst of recent vegetation. In Rhaetic, Jurassic and Wealden floras, the Ginkgoales were exceedingly abundant (Map B, G₁–G₁₇); in addition to leaves agreeing almost exactly with those of the recent Ginkgoales. species (fig. 18), there are others separated as a distinct genus, Baiera (fig. 18, G), characterized by the greater number and narrower form of the segments, which may be best compared with such leaves as those of the recent fern Actiniopteris and of certain species of Schizaea. Male flowers, like those of Ginkgo biloba, but usually characterized by a rather larger number of oval pollen-sacs on the stamens, have been found in England, Germany, Siberia and elsewhere in association with Ginkgo and Baiera foliage. The occasional occurrence of three or even four pollen-sacs on the stamens of the recent species affords a still closer agreement between the extinct and living types. Seeds like those of Ginkgo biloba have also been recorded as fossils in Jurassic rocks, and it is possible that the type of flower known as Beania, from the Inferior Oolite rocks of Yorkshire, may have been borne by Ginkgo or Baiera. The regions from which satisfactory examples of Ginkgoales (Baiera or Ginkgo) have been recorded are shown in Map B (G₁–G₁₇). Both Tertiary and Mesozoic localities are indicated in the map.

Fig. 18.—Leaves of Ginkgoales.

  A, Ginkgodium, Japan (Jurassic).

  B, C, D, E, F, H, Ginkgo leaves.—B, from Franz Josef Land (Jurassic);
   C, Greenland (Lower Cretaceous); D, Siberia (Jurassic);
   E, Germany (Wealden); F, England (Jurassic); H, China (Rhaetic).

  G, Baiera leaf, Inferior Oolite, England.

  (A, after Yokoyama; B, after Nathorst; C, D, after Heer; E. after Schenk;
   H, after Krasser. All the figures 1/2 nat. size.)

An adequate account of fossil Mesozoic Conifers is impossible within the limits of this article. Coniferous twigs are very common in Mesozoic strata, but in most cases we are compelled to refer them to provisional genera, as the evidence of vegetative shoots alone is not sufficient to enable us to determine their position within the Coniferae. There are, however, Coniferales. several forms which it is reasonable to include in the Araucarieae; that this family was to the fore in the vegetation of the Jurassic period is unquestionable. We have not merely the striking resemblance of vegetative shoots to those of recent species of Araucaria and Agathis, e.g. species of Nageiopsis, abundantly represented in the Upper Jurassic beds of the Potomac area in North America, species of Pagiophyllum and other genera of Jurassic and Wealden age, but an abundance of fossil wood (Araucarioxylon) from Jurassic and Cretaceous strata in Europe, North America, Madagascar and elsewhere agreeing with that of recent Araucarieae, in addition to several well-preserved female flowers. C. A. Hollick and E. C. Jeffrey have recently shown that some Lower Cretaceous specimens of the well-known genus Brachyphyllum obtained from Staten Island, N.Y., possess wood of the Araucarian type. This genus has long been known as a common and widely spread Jurassic and Cretaceous conifer, but owing to the absence of petrified specimens and of well-preserved cones, it has been impossible to refer it to a definite position in the Coniferales. It is now clear that some at least of the species of Brachyphyllum must be referred to the Araucarieae. In a recently published paper Seward and Ford have given a general account of the Araucarieae, recent and extinct, to which reference may be made for further details as to the geological history of this ancient section of the Coniferales. Some of the fossils referred to the genus Kaidocarpon, and originally described as monocotyledonous inflorescences, are undoubted Araucarian cones; other cones of the same type have been placed in the genus Cycadeostrobus and referred to Cycads. Araucarites Hudlestoni, described by Mr Carruthers from the Coralline Oolite rocks of Malton in Yorkshire; Araucarites sphaerocarpa from the Inferior Oolite of Somerset; also another cone found in the Northampton Sands, which is probably specifically identical with A. Hudlestoni, and named by Carruthers Kaidocarpon ooliticum, afford good illustrations of British Araucarian flowers. A flower of a rather different type, Pseudaraucaria major, exhibiting in the occurrence of two seeds in each scale an approach to the cones of Abietineae, has been described by Professor Fliche from Lower Cretaceous rocks of Argonne. The well-known Whitby jet of Upper Liassic age appears to have been formed to a large extent from Araucarian wood. Among the more abundant Conifers of Jurassic age may be mentioned such genera as Thuytes and Cupressites, which agree in their vegetative characters with members of the Cupressineae, but our knowledge of the cones is far from satisfactory. Many of the small female flowers borne on shoots with foliage of the Cupressus type consist of spirally disposed and not verticillate scales, e.g. Thuytes expansus, a common Jurassic species.

Fossil wood, described under the name Cupressinoxylon, has been recorded from several Mesozoic horizons in Europe and elsewhere, but this term has been employed in a wide sense as a designation for a type of structure met with not only in the Cupressineae, but in members of other families of Coniferae. The Abietineae do not appear to have played a prominent part before the Wealden period; various older species, e.g. Rhaetic specimens from Scania, are recorded, but it is not until we come to the Upper Jurassic and Wealden periods that this modern family was abundantly represented. Fossil wood of the Pinites type (Pityoxylon) has been described from England, France, Germany, Sweden, Spitsbergen, North America and elsewhere; some of the best British examples have been obtained from the so-called Pine-raft, the remains of water-logged and petrified wood of Lower Greensand age, seen at low water near Brook Point in the Isle of Wight. Well-preserved Abietineous female flowers have been obtained from the Wealden rocks of England and Belgium, e.g. Pinites Dunkeri, P. Solmsi, &c.; specimens of seeds and vegetative shoots are recorded also from Spitsbergen and other regions. Hollick and Jeffrey have recently added to our knowledge of the anatomy of Cretaceous species of Pinus, and Miss Stopes and Dr Fujii have made important contributions on the structure of Cretaceous plants from Japan. Cones of Lower Cretaceous age have been described by Fliche from Argonne, which bear a close resemblance to the female flowers of recent species of Cedrus. The two surviving species of Sequoia afford an illustration of the persistence of an old type, but unfortunately most of the Mesozoic species referred to this genus do not possess sufficiently perfect cones to confirm their identification as examples of Sequoia. Some of the best examples of cones and twigs referred to Sequoia are those described by Heer from Cretaceous rocks of Greenland, and Professor D. P. Penhallow of Montreal has described the anatomical structure of the stem of Sequoia Langsdorfii, a Tertiary species occurring in Europe and North America.

There are a few points suggested by a general survey of the Mesozoic floras, which may be briefly touched on in conclusion. In following the progress of plant-life through those periods in the history of the earth of which records are left in ancient sediments, seams of coal or old land-surfaces, we recognize at certain stages a want of continuity between the floras of successive ages. The imperfection of the geological record, considered from the point of view of evolution, has been rendered familiar by Darwin’s remarkable chapter in the Origin of Species. Breaks in the chain of life, as represented by gaps in the blurred and incomplete documents afforded by fragmentary fossils, are a necessary consequence of the general plan of geological evolution; they mark missing chapters rather than sudden breaks in an evolutionary series. On the other hand, a study of the plant-life of past ages tends to the conviction that too much stress may be laid on the imperfection of the geological record as a factor in the interpretation of palaeontological data. The doctrine of Uniformitarianism, as propounded by Lyell, served to establish geology on a firmer and more rational basis than it had previously possessed; but latterly the tendency has been to modify the Lyellian view by an admission of the probability of a more intense action of groups of forces at certain stages of the earth’s history. As a definite instance a short review may be given of the evidence of palaeobotanical records as regards their bearing on plant-evolution. Starting with the Permo-Carboniferous vegetation, and omitting for the moment the Glossopteris flora, we find a comparatively homogeneous flora of wide geographical range, consisting to a large extent of arborescent lycopods, calamites, and other vascular cryptogams, plants which occupied a place comparable with that of Gymnosperms and Angiosperms in our modern forests; with these were other types of the greatest phylogenetic importance, which serve as finger-posts pointing to lines of evolution of which we have but the faintest signs among existing plants. Other types, again, which may be referred to the Gymnosperms, played a not unimportant part in the Palaeozoic vegetation. No conclusive proof has so far been adduced of the existence in those days of the Cycads, nor is there more than partial evidence of the occurrence of genera which can be placed with confidence in any of the existing families of Conifers. There are, moreover, no facts furnished by fossil plants in support of the view that Angiosperms were represented either in the low-lying forests or on the slopes of the mountains of the Coal period. Passing higher up the geological series, we find but scanty records of the vegetation that existed during the closing ages of the Permian period, and of the plants which witnessed the beginning of the Triassic period we have to be content with the most fragmentary relics. It is in rocks of Upper Triassic and Rhaetic age that abundant remains of rich floras are met with, and an examination of the general features of the vegetation reveals a striking contrast between the Lower Mesozoic plants and those of the Palaeozoic period. Arborescent Pteridophytes are barely represented, and such dominant types as Lepidodendron, Sigillaria, Calamites and Sphenophyllum have practically ceased to exist; Cycads and Conifers have assumed the leading role, and the still luxuriant fern vegetation has put on a different aspect. This description applies almost equally to the floras of the succeeding Jurassic and Wealden periods. The change to this newer type of vegetation was no doubt less sudden than it appears as read from palaeobotanical records, but the transition period between the Palaeozoic type of vegetation and that which flourished in the Lower Mesozoic era, and continued to the close of the Wealden age, was probably characterized by rapid or almost sudden changes. In the southern hemisphere the Glossopteris flora succeeded a Lower Carboniferous vegetation with a rapidity similar to that which marked the passage in the north from Palaeozoic to Mesozoic floras. This apparently rapid alteration in the character of the southern vegetation took place at an earlier period than that which witnessed the transformation in the northern hemisphere. The appearance of a new type of vegetation in India and the southern hemisphere was probably connected with a widespread lowering of temperature, to which reference has already been made. It was from this Glossopteris flora that several types gradually migrated across the equator, where they formed part of the vegetation of more northern regions. The difference between the Glossopteris flora and those which have left traces in the Upper Gondwana rocks of India, in the Wianamatta and Hawkesbury beds of Australia, and in the Stormberg series of South Africa is much less marked than that between the Permo-Carboniferous flora of the northern hemisphere and the succeeding Mesozoic vegetation. In other words, the change took place at an earlier period in the south than in the north. To return to the northern hemisphere, it is clear that the Wealden flora, as represented by plants recorded from England, France, Belgium, Portugal, Russia, Germany and other European regions, as also from Japan and elsewhere, carries on, with minor differences, the facies of the older Jurassic floras. It was at the close of the Wealden period that a second evolutionary wave swept over the vegetation of the world. This change is most strikingly illustrated by the inrush of Angiosperms, in the equally marked decrease in the Cycads, and in the altered character of the ferns. It would appear that in this case the new influence, supplied by the advent of Angiosperms, had its origin in the north. Unfortunately, our knowledge of the later floras in the southern hemisphere is very incomplete, but a similar transformation appears to have characterized the vegetation south of the equator. As to the nature of the chief factors concerned in the two revolutions in the vegetable kingdom, if it is admissible to use so strong a term, only a guess can be hazarded. Physical conditions no doubt played an important part, but whatever cause may have had the greatest share in disturbing the equilibrium of evolutionary forces, it would seem that the apparently sudden appearance of Cycads and other types at the close of the Palaeozoic period made a widespread and sudden impression on the whole character of the vegetation. At a later stage—in post-Wealden days—it was the appearance of Angiosperms, probably in northern latitudes, that formed the chief motive power in accelerating the transition in the facies of plant-life from that which marked what we have called the Mesozoic floras, to the vegetation of the Upper Cretaceous and Tertiary periods. With the advent of Angiosperms began, as the late marquis of Saporta expressed it, “Une révolution, ainsi rapide dans sa marche qu’universelle dans ses effets.” From the floras of the Tertiary age we pass by gradual stages to those which characterize the present phase of evolutionary progress. Among modern floras we find here and there isolated types, such as Ginkgo, Sequoia, Matonia, Dipteris and the Cycads, persisting as more successful survivals which have held their own through the course of ages; these plants remain as vestiges from a remote past, and as links connecting the vegetation of to-day with that of the Mesozoic era.

Authorities.—Glossopteris Flora: Blanford, H. F., “On the age and correlation of the Plant-bearing Series of India, &c.,” Quarterly Journal Geol. Soc. xxxi. (1875); Feistmantel, “Fossil Flora of the Gondwana System,” Mem. Geol. Surv. India, vols. iii., &c. (1879, &c.); Seward, Fossil Plants as Tests of Climate (Cambridge, 1892), with bibliography; “The Glossopteris Flora,” Science Progress, with bibliography; “On the Association of Sigillaria and Glossopteris in South Africa,” Q.J.G.S., vol. liii. (1897); E. A. N. Arber, Catalogue of the Fossil Plants of the Glossopteris Flora in the Department of Geology (British Museum, Nat. Hist., Brit. Mus. Catalogue (London, 1905), with full bibliography; Medlicott and Blanford, Manual of the Geology of India (2nd ed., Oldham, R. D., Calcutta, 1893); David, “Evidences of Glacial Action in Australia in Permo-Carboniferous time,” Q.J.G.S., vol. lii. (1896); Zeiller, Éléments de paléobotanique (Paris, 1900); Potonié, “Fossile Pflanzen aus deutsch und portugiesisch Ostafrika,” Deutsch-Ostafrika, vii. (Berlin, 1900), with bibliography. General: Potonié, Lehrbuch der Pflanzenpalaeontologie (Berlin, 1899); Scott, Studies in Fossil Botany (1900); Seward, Fossil Plants (Cambridge: vol. i., 1898); vol. ii. 1910, with bibliography; Zeiller, “Revue des travaux de paléontologie végétale,” Rev. gén. bot. (1903) et seq. Catalogue of the Mesozoic Plants in the British Museum, (a) “Wealden Flora,” pts. i. and ii.; (b) “Jurassic Flora,” pt. i. (1894–1901), pt. ii. (1904), with bibliography; “On the Structure and Affinities of Matonia pectinata, with Notes on the Geological History of the Matonineae,” Phil. Trans. cxci. (1899); “On the Structure, &c., of Dipteris,” ibid. cxciv. (1901, with bibliography; Seward and Ford, “The Araucarieae, recent and extinct,” Phil. Trans. R. Soc. (London, 1906); G. R. Wieland, “American Fossil Cycads,” Publication Carnegie Instit. (Washington, 1906); Nathorst, “Paläobotanische Mitteil.,” K. Svensk. Vetenskaps. Akad. Hand. xlii., No. 5 (1907); The Norwegian North-Polar Expedition, iii. (1893–1896); “Fossil Plants from Franz Josef Land;” L. F. Ward, “Status of the Mesozoic Floras of the United States,” Twentieth Ann. Rep. Geol. Survey (Washington, 1900); Solnis-Laubach, “Ueber das Genus Pleuromeia,” Bol. Zeit. (1899); Newton and Teall, “Notes on a Collection of Rocks and Fossils from Franz Josef Land,” Q.J.G.S. liii. (1897); Hollick and Jeffrey, “Studies of Cretaceous Coniferous remains,” Mem. New York Botanical Garden, vol. iii. (1909); Stopes and Fujii, “Structure and Affinities of Cretaceous Plants,” Phil. Trans. R. Soc. (1910). References to important papers on Mesozoic botany will be found in the bibliographies mentioned in the above list.  (A. C. Se.) 

III.—Tertiary

After the Wealden period, and before the deposition of the lowest strata of the Chalk, so remarkable a change takes place in the character of the vegetation that this break must be taken as, botanically, the transition point from a Secondary to a Tertiary flora. A flora consisting entirely, with a single doubtful exception, of Lower Cretaceous. Gymnosperms and Cryptogams gives place to one containing many flowering plants; and these increase so rapidly that before long they seem to have crowded out many of the earlier types, and to have themselves become the dominant forms. Not only do Angiosperms suddenly become dominant in all known plant-bearing deposits of Upper Cretaceous age, but strangely enough the earliest found seem to belong to living orders, and commonly have been referred to existing genera. From Cretaceous times onwards local distribution may change; yet the successive floras can be analysed in the same way as, and compared with, the living floras of different regions. World-wide floras, such as seem to characterize some of the older periods, have ceased to be, and plants are distributed more markedly according to geographical provinces and in climatic zones. This being the case, it will be most convenient to discuss the Tertiary floras in successive order of appearance, since the main interest no longer lies in the occurrence of strange extinct plants or of transitional forms connecting orders now completely isolated.

 

Fig. 1.—Alismacites primaevus.

The accurate correlation in time of the various scattered plant-bearing deposits is a matter of considerable difficulty, for plant-remains are preserved principally in lacustrine strata laid down in separate basins of small extent. This it is obvious must commonly be the case, as most leaves and fruits are not calculated to drift far in the sea without injury or in abundance; nor are they likely as a rule to be associated with marine organisms. Deposits containing marine fossils can be compared even when widely separated, for the ocean is continuous and many marine species are world-wide. Plants, on the other hand, like land and fresh-water animals, occupied areas which may or may not have been continuous. Therefore, without a knowledge of the physical geography of any particular period, we cannot know whether like or unlike floras might be expected in neighbouring areas during that period. If, however, we discover plant-bearing strata interstratified with deposits containing marine fossils, we can fix the period to which the plants belong, and may be able to correlate them in distinct areas, even though the floras be unlike. This clear stratigraphical evidence is, however, so rarely found that much uncertainty still remains as to the true age of several of the floras now to be described.

In rocks approximately equivalent to the Lower Greensand of England, or slightly earlier, Angiosperms make their first appearance; but as the only strata of this age in Britain are of marine origin, we have to turn to other countries for the evidence. The earliest Angiosperm yet found in Europe is a single monocotyledonous leaf of doubtful affinities, named by Saporta Alismacites primaevus (fig. 1), and found in the Valenginian strata of Portugal. These deposits seem to be equivalent to British Wealden rocks, though in the latter, even in their upper part, no trace of Angiosperms has been discovered. No other undoubted Angiosperm has yet been discovered in Europe in strata of this age, but Heer records a poplar-like leaf from Urgonian strata, a stage newer than the Valenginian, in Greenland, and Saporta has described from strata of the same date in Portugal a Euphorbiaceous plant apparently closely allied to the living Phyllanthus and named by him Choffatia Francheti (fig. 2). We must turn to North America for a fuller knowledge of the earliest flowering-plants.

In S. Dakota a remarkable series has been discovered, lying unmistakably between marine Upper Jurassic rocks below and Upper Cretaceous above. There has been a certain amount of confusion as to the exact strata in which the plants occur, but this has now been cleared up by the researches of Lester F. Ward, who has shown how the American Cretaceous. Secondary flora gives place to one of Tertiary character.

 

Fig. 2.—Choffatia Francheti.

The lower strata—i.e. those most allied to the Jurassic—contain only Gymnosperms and Cryptogams. The next division (Dakota No. 2 of Meek and Hayden) contains Gymnosperms and Ferns of Neocomian types, or even of Neocomian species; but mingled with these occur a few dicotyledonous leaves belonging to four genera. The specimens are very fragmentary, and all that can be said is that one of the forms may be allied to oak, another to fig, a third to Sapindus, and the fourth may perhaps be near to elm. The “Potomac Formation” of Virginia and Maryland is doubtless also mainly of Neocomian age, for though it rests unconformably on much older strata, the successive floras found in it are so allied to those of S. Dakota as to leave little doubt as to the general homotaxis of the series. Lester Ward records no fewer than 737 distinct forms, consisting chiefly of Ferns, Cycads, Conifers and Dicotyledons, the Ferns and Cycads being confined mainly to the Older Potomac, while the Dicotyledons are principally represented in the Newer Potomac, though occurring more rarely even down to the base of the series. Six successive stages have been defined in the Potomac formation. The Mount Vernon beds, which occur about the middle of the series, have as yet yielded only a small number of species, though these include the most interesting early Angiosperms. Among them are recorded a Casuarina, a leaf of Sagittaria (which however, as observed by Zeiller, may belong to Smilax), two species of poplar-like leaves with remarkably cordate bases, Menispermites (possibly a water-lily) and Celastrophyllum (perhaps allied to Celastrus). Proteophyllum, found in the same bed, and also in the Infra-Cretaceous of Portugal, seems to have belonged to a Proteaceous plant, though only leaves without fruits have yet been discovered in deposits of this early date. Whatever doubt may be left as to the exact botanical position of these early Lower Cretaceous Angiosperms, it is clear that both Monocotyledons and Dicotyledons are represented by several types of leaves, and that the flora extended over wide areas in North America and Greenland, and is found again at a few points in Europe. There is yet no clear evidence either of climatic zones or of the existence of geographical provinces during this period.

The next strata, the Aquila Creek series, contain a well-marked dicotyledonous flora, in which both the form and nervation of the leaves begin to approximate to those of recent times. The leading characteristic of this Middle Potomac flora is the proportion of Dicotyledons. Notwithstanding this apparent passage-bed, there is a marked difference between the Older and the Newer Potomac floras, very few species passing from the one to the other. Only 15 out of 405 plants in the older series occur in the beds above, though already more than 350 species have been determined from this newer series. The plants from the Amboy Clays, which form the most important division of the Newer Potomac series and were monographed in 1895 by J. S. Newberry, seem to belong to the commencement of the Upper Cretaceous period. It is remarkable that nearly 80% of the species are Dicotyledons, and that no Monocotyledons have been found. The mere enumeration of the genera will indicate how close the flowering plants are to living forms. Newberry records Juglans, Myrica (7 species), Populus, Salix (5 species), Quercus, Planera, Ficus (3 species), Persoonia and another extinct Proteaceous genus named Proteoides, Magnolia (7 species), Liriodendron (4 species), Menispermites, Laurus and allied plants, Sassafras (3 species), Cinnamomum, Prunus, Hymenaea, Dalbergia, Bauhinia, Caesalpinia, Fontainea, Colutea and other Leguminosae, Ilex, Celastrus, Celastrophyllum (10 species), Acer, Rhamnites, Paliurus, Cissites, Tiliaephyllum, Passiflora, Eucalyptus (5 species), Hedera, Aralia (8 species), Cornophyllum, Andromeda (4 species), Myrsine, Sapotacites, Diospyros, Acerates, Viburnum and various genera of uncertain affinities. The points that suggest themselves with regard to this flora are, that it includes a fair representation of the existing orders of warm-temperate deciduous trees; that the more primitive types—such as the Amentaceae—do not appear to preponderate to a greater extent than they do in the existing temperate flora; that the assemblage somewhat suggests American affinities; and that when we take into account deficient collecting, local conditions, and the non-preservation of succulent plants, there is no reason for saying that certain other orders must have been absent. The great rarity of Monocotyledons is a common characteristic of fossil floras known only, as this one is, from leaves principally belonging to deciduous trees. With regard to suggested American affinities, it must be borne in mind that the Neocomian Angiosperms are little known except in America and in Greenland, and that we therefore cannot yet say whether families now mainly American were not formerly of world-wide distribution. We know that this was the case with some, such as Liriodendron; and in Eucalyptus we see the converse, where a genus formerly American is now confined to a far distant region. The Neocomian flora has been collected from an area extending over about 30° of latitude; but there is little evidence of any corresponding climatic change. We cannot yet say, however, that the deposits are exactly contemporaneous, and the great climatic variations that have taken place in the northern hemisphere during the existence of our living flora should make us hesitate to correlate too minutely from the evidence of plants alone.

The highest division of the Dakota series (known as Dakota No. 1) which lies immediately beneath Upper Cretaceous strata with marine fossils, contains a flora so like that of the Tertiary deposits that only the clearest geological evidence has been considered sufficient to prove that Heer was wrong when he spoke of the plants as Miocene. These highest plant-bearing strata rest, according to Lester Ward, somewhat unconformable on the Dakota No. 2; they show also a marked difference in the included plants. The genera of Dicotyledons represented are Quercus, Sassafras, Platanus, Celastrophyllum, Cissites, Viburnites.

In the central parts of North America the lacustrine plant-bearing deposits are of enormous thickness, the Dakota series being followed by marine Cretaceous strata known as the Colorado and Montana groups, and these being succeeded conformably by a thousand feet or more of lacustrine shales, sandstones and coal-seams, belonging to the Laramie series. This also contains occasional marine Upper Cretaceous fossils, as well as reptiles of Cretaceous types. An extensive literature has grown up relating to these Laramie strata, for owing to the Tertiary aspect of the contained plants, geologists were slow to recognize that they could be truly contemporaneous and interbedded with others yielding Cretaceous animals. In addition to this, the earlier writers included in the Laramie series many deposits now known to be of later date and truly Tertiary, and the process of separation is even now only partially completed. It will be safest in these circumstances to accept as our guide to the true Laramie flora the carefully compiled “Catalogue” of F. H. Knowlton. According to this catalogue, the true Laramie flora includes about 250 species, more than half of which are deciduous forest trees, herbaceous Dicotyledons, Monocotyledons and Cryptogams, all being but poorly represented. Among the few Monocotyledons are leaves and fruits of palms, and traces of grasses and sedges. The Dicotyledons include several water-lilies, a somewhat doubtful Trapa, and many genera of forest trees still common in America. The genera best represented are Ficus (21 species), Quercus (16 species), Populus (11 species), Rhamnus (9 species), Platanus (8 species), Viburnum (7 species), Magnolia (6 species), Cornus (5 species), Cinnamomum (5 species), Juglans (4 species), Acer (4 species), Salix (4 species), Aralia (3 species), Rhus (3 species), Sequoia (3 species). Of trees now extinct in America, Eucalyptus and Ginkgo are perhaps the most noticeable. So large a proportion of the trees still belongs to the flora of North America that one is apt to overlook the fact that among the more specialized plants some of the largest American orders, such as the Compositae, are still missing from strata belonging to the Cretaceous period.

The imperfection and want of continuity of the records in Europe have made it necessary in dealing with the Cretaceous floras for us to give the first place to America. But it is now advisable to return to Europe, where Upper Cretaceous plants are not uncommon, and the position of the deposits in the Cretaceous series can often European Cretaceous. be fixed accurately by their close association with marine strata belonging to definite subdivisions. As these divisions of Cretaceous time will have to be referred to more than once, it will be useful to tabulate them, thus showing which plant-beds seem to be referable to each, and what are the British strata of like age. It has not yet been found possible so closely to correlate the strata of Europe with those of America, where distance has allowed geographical differences in both fauna and flora to come into play; therefore, beyond the references to Lower or Upper Cretaceous, no classification of the American Cretaceous strata has here been given. In Europe the most commonly accepted divisions of the Cretaceous period are as follows:—

England.

France, &c.

 Wanting

 Danian

 Upper Chalk

 Senonian

 Middle Chalk

 Turonian

Lower Chalk	$\scriptstyle {\left.{\begin{matrix}\ \\\ \end{matrix}}\right\}\,}$
Upper Green-sand

 Cenomanian

 Gault

 Albian

 Aptian

 Lower Green-sand

 Valenginian

 Urgonian

 Wealden

 Neocomian

In the continental classification the deposits from the Gault downwards are grouped as Lower Cretaceous; but in Great Britain there is a strong break below the Gault and none above; and the Gault is therefore classed as Upper Cretaceous. The limits of the divisions in other places do not correspond, the British and continental strata often being so unlike that it is almost impossible to compare them. The doubt as to the exact British equivalent of the Valenginian strata of Portugal, which yield the earliest Dicotyledon, has already been alluded to. The plant-bearing deposits next in age, which have yielded Angiosperms, appear to belong to the Cenomanian, though from Westphalia a few species belonging to the Cryptogams and Gymnosperms, found in deposits correlated with the Gault, have been described by Hosius and von der Marck.

Fig. 3.—Credneria triacuminata.

In Great Britain the whole of the Upper Cretaceous strata are of marine origin, and have yielded no land-plants beyond a few fir-cones, drift-wood and rare Dicotyledonous leaves in the Lower Chalk. Most of the deposits which have yielded Angiosperms of Cretaceous age in central Europe correspond in age with the English Upper Chalk (Senonian), but a small Cenomanian flora has been collected from the Unter Quader in Moravia. Heer described from this deposit at Moletein 13 genera, of which 7 are still living, containing 18 species, viz.: 1 fern, 4 Conifers, 1 palm, 2 figs, 1 Credneria, 2 laurels, 1 Aralia, 1 Chondrophyllum (of uncertain affinities), 2 magnolias, 2 species of Myrtaceae and a species of walnut. Saxony yields from strata of this period at Niederschoena 42 species, described by Ettingshausen. This small flora is most remarkable, for no fewer than 6 genera, containing 8 species, are referred to the Proteaceae. The Cenomanian flora of Bohemia is larger and equally peculiar. Among the Dicotyledons described by Velenovsky are the following: Credneria (5 species), Araliaceae (17 species), Proteaceae (8 species), Myrica (2 species), Ficus (5 species), Quercus (2 species), Magnoliaceae (5 species), Bombaceae (3 species), Laurineae (2 species), Ebenaceae (2 species), Verbenaceae, Combretaceae, Sapindaceae (2 species), Camelliaceae, Ampelideae, Mimoseae, Caesalpinieae (5 species), Eucalyptus (2 species), Pisonia, Phillyrea, Rhus, Prunus, Bignonia, Laurus, Salix, Benthamia. To this list Bayer adds Aristolochia. The Cenomanian flora of central Europe appears to be a subtropical one, with marked approaches to the living flora of Australia. The majority of its Dicotyledons belong to existing genera, but one of the most prolific and characteristic Cretaceous forms is Credneria (Fig. 3), a genus of doubtful affinities, which has been compared by different authors to the poplars, planes, limes and other orders.

The Cretaceous plant-beds of Westphalia include both Upper and Lower Senonian, the two floras being very distinct. Hosius and von der Marck describe, for instance, 12 species of oak from the Upper and 6 from the Lower strata, but no species is common to the two. The same occurs with the figs, with 3 species above and 8 below. The 6 species of Credneria are all confined to the older deposits. In fact, not a single Dicotyledon is common to these two closely allied divisions of the Cretaceous series; a circumstance not easy to explain, when we see how well the oaks and figs are represented in each. Four species of Dewalquea, a ranunculaceous genus allied to the hellebore, make their appearance in the Upper Senonian of Westphalia, other species occurring at Aix-la-Chapelle in deposits of about the same age. The Senonian flora of the last-named place, and that of Maestricht, are still only imperfectly known. It is unnecessary to trace the variations of the Upper Cretaceous flora from point to point; but the discoveries within the Arctic circle have been so surprising that attention must again be called to them. Besides the Lower Cretaceous plants already mentioned, Heer has described from Greenland a flora of Cenomanian age, and another belonging to the Senonian. The Cenomanian strata have yielded already 177 species, the different groups being represented in these proportions: Cryptogams, 37, 30 of which are Ferns; Cycads, 8; Conifers, 27; Monocotyledons, 8; Apetalous Dicotyledons, 31; other Dicotyledons, 66. The Senonian strata have yielded 118 species, 21 of which are Cryptogams, 11 Conifers, 5 Monocotyledons, 75 Dicotyledons. Forest trees, especially oaks, are plentiful, and many of the species are identical with those found in Cretaceous deposits in more southern latitudes. Both of these floras suggest, however, that the climate of Greenland was somewhat colder than that of Westphalia, though scarcely colder than warm-temperate.

The Cretaceous deposits just described are followed by a series of Tertiary formations, but in Europe the continuity between Cretaceous and Tertiary is not quite complete. The Tertiary formations have been assigned to six periods; these are termed—Paleocene, Eocene, Oligocene, Miocene, Pliocene, Pleistocene, and each has its own botanical peculiarities.

During the Paleocene period the plants were not markedly different from those of the Upper Cretaceous. Its flora is still but imperfectly known, for we are dependent on two or three localities for the plants. There is found at Sézanne, about 60 m. east of Paris, an isolated deposit of calcareous tufa full of leaves, which gives a curious Paleocene Plants. insight into the vegetation which flourished in Paleocene times around a waterfall. Sézanne yields Ferns in profusion, mingled with other shade-loving plants such as would grow under the trees in a moist ravine; its vegetation is comparable to that of an island in the tropical seas. Monocotyledons are rare, the only ones of much interest being some fragments of pandanaceous leaves. The absence of Gymnosperms is noticeable. The Proteaceae are also missing; but other Dicotyledons occur in profusion, many of them being remarkable for the large size of their deciduous leaves. Among the flowering plants are Dewalquea, a ranunculaceous genus already mentioned as occurring in the Upper Cretaceous, and numerous living genera of forest-trees, such as occur throughout the Tertiary period, and are readily comparable with living forms. Saporta has described about seventy Dicotyledons, most of which are peculiar to this locality.

The plant-bearing marls of Gelinden, near Liège, contain the débris of a Paleocene forest. The trees seemed to have flourished on neighbouring chalky heights. The most abundant species of this forest were the oaks and chestnuts, of which a dozen have been collected; laurels, Viburnum, ivy, several Aralias, Dewalquea, a Thuja and several Ferns may be added. This flora is compared by Saporta and Marion with that of southern Japan. Other deposits of this age in France have furnished plants of a more varied aspect, including myrtles, araucarias, a bamboo and several fan-leaved palms. Saporta points out the presence in these Paleocene deposits of certain types common, on the one hand, to the American Tertiary strata between the Missouri and the Rocky Mountains, and on the other, to the Tertiary flora of Greenland. The Paleocene deposits of Great Britain are of marine origin, and only yield pine-cones and fragments of Osmunda.

The British Eocene and Oligocene strata yield so large a flora, and contain plant-beds belonging to so many different stages, that it is unfortunate we have still no monograph on the subject, the one commenced by Ettingshausen and Gardner in 1879 having reached no farther than the Ferns and Gymnosperms. This deficiency Eocene and Oligocene of Great Britain. makes it impossible to deal adequately with the British Eocene plants, most of the material being either unpublished or needing re-examination.

In the earliest Eocene plant-beds, in the Woolwich and Reading series, a small but interesting flora is found, which suggests a temperate climate less warm than that of earlier or of later periods. Leaves of planes are abundant, and among the plants recorded are two figs, a laurel, a Robinia, a Grevillea and a palm. Ferns are scarce, Ettingshausen and Gardner recording only Aneimia subcretacea and Pteris (?) Prestwichii. The only Gymnosperms determined are Libocedrus adpressa, which is close to L. decurrens of the Yosemite, and Taxodium europaeum. A few plants have been found in the next stage, the Oldhaven beds, and among these are fig and cinnamon. Gardner considers the plants to point to subtropical conditions. The London Clay has yielded a large number of plants, but most of the species are represented by fruits alone, not by leaves. This circumstance makes it difficult to compare the flora with that of other formations, for not only is it uncertain which leaves and fruits belong to the same plant, but there is the additional source of doubt, that different elements of the same flora may be represented at different localities. Of some plants only the deciduous leaves are likely to be preserved, whilst other succulent-leaved forms will only be known from their woody fruits. Among the 200 plants of the London Clay are no Ferns, but 6 genera of Gymnosperms—viz. Callitris (2 species), Sequoia, Athrotaxis (?) Ginkgo, Podocarpus, Pinus; and several genera of palms, of which the tropical Nipa is the most abundant and most characteristic, among the others being fan-palms of the genera Sabal and Chamaerops. The Dicotyledons need further study. Among the fruits Ettingshausen records Quercus, Liquidambar, Laurus, Nyssa, Diospyros, Symplocos, Magnolia, Victoria, Hightea, Sapindus, Cupania, Eugenia, Eucalyptus, Amygdalus; he suggests that the fruits of the London Clay of Sheppey may belong to the same plants as the leaves found at Alum Bay in the Isle of Wight.

The next stage is represented by the Lower Bagshot leaf-beds of Alum Bay. These pipeclays yield a varied flora, Ettingshausen recording 274 species, belonging to 116 genera and 63 families. Gardner, however, is unable to reconcile this estimated richness with our knowledge of the flora, and surmises that fossil plants from other localities must have been inadvertently included. He considers the flora to be the most tropical of any that has so far been studied in the northern hemisphere. Its most conspicuous plants are Ficus Bowerbankii, Aralia primigenia, Comptonia acutiloba, Dryandra Bunburyi, Cassia Ungeri and the fruits of Caesalpinia. The floras which it chiefly resembles are first, that of Monte Bolca, and second, that of the Gres du Soissonais, which latter Gardner thinks may be of the same age, and not earlier, as is generally supposed. The total number of species found at Alum Bay, according to this author, is only about 50 or 60.

To the Bagshot Sand succeeds the thick mass of sands with intercalated plant-beds seen in Bournemouth cliffs. Each bed yields peculiar forms, the total number of species amounting to many hundred, most of them differing from those occurring in the strata below. The plants suggest a comparison of the climate and forests with those of the Malay Archipelago and tropical America. At one place we find drifted fruits of Nipa, at another Hightea and Anona. Other beds yield principally palms, willows, laurels, Eucalyptus or Ferns; but there are no Cycads. As showing the richness of this flora, we may mention that in the only orders which have yet been monographed, Ferns are represented by 17 species and Gymnosperms by 10, though these are not the groups best represented. Gardner speaks of the Bournemouth flora as appearing to consist principally of trees or hard-wooded shrubs, comparatively few remains of the herbaceous vegetation being preserved. The higher Eocene strata of England—those above the Bournemouth Beds—are of marine origin, and yield only drifted fruits, principally fir-cones.

In the volcanic districts of the south-west of Scotland and the north-east of Ireland plant-beds are found intercalated between the lava-flows. These also, like the lignites of Bovey Tracey, have been referred to the Miocene period, on the supposed evidence of the plants; but more recent discoveries by Gardner tend to throw doubt on this allocation, and suggest that, though of various ages, the first-formed of these deposits may date back to early Eocene times. The flora found in Mull points distinctly to temperate conditions; but it is not yet clear whether this indicates a different period from the subtropical flora of the south of England, or whether the difference depends on latitude or local conditions. The plants include a Fern, Onoclea hebridica, close to a living American form; four Gymnosperms belonging to the genera Cryptomeria, Ginkgo, Taxus and Podocarpus; Dicotyledons of about 30 species, several of which have been figured. Among the Dicotyledons may be mentioned Platunus, Acer (?), Quercus (?), Viburnum, Alnus, Magnolia, Corylus (?), Castanea (?), Zizyphus, Populus and the nettle-like Boehmeria antiqua. The absence of the so-called cinnamon-leaves and the Smilaceae, which always enter into the composition of Middle Eocene and Oligocene floras, is noticeable. The Irish strata yield two ferns; 7 Gymnosperms, Cupressus, Cryptomeria, Taxus, Podocarpus, Pinus (2 species), Tsuga; and leaves of about 25 Dicotyledons. The most abundant leaf, according to Gardner, does not seem distinct from Celastrophyllum Benedeni, of the Paleocene strata of Gelinden; a water-lily, Nelumbium Buchii, occurs also in Oligocene beds on the Continent; the species of MacClintockia (fig. 4) is found both in the Arctic floras and at Gelinden. Among the other plants are an alder, an oak and a doubtful cinnamon.

Fig. 4.—MacClintockia trinervata.

Leaving these Scottish and Irish deposits of doubtful age, we find in the Hampshire Basin a thick series of fluviatile, lacustrine and marine deposits undoubtedly of Lower and Middle Oligocene date. Their flora is still a singularly poor one, though plants have been obtained at many different levels; they perhaps indicate a somewhat cooler climate than that of the Bournemouth series. Among the more abundant plants are nucules of several species of Chara, and drifted fruits and seeds of water-lilies, of Folliculites (now generally referred to Stratiotes) and of Limnocarpus (allied to Potamogeton); there is little else mixed with these. Other seams are full of the twigs and cones of Athrotaxis, a Conifer now confined to Tasmania. Ferns are represented by Gleichenia, Lygodium and Chrysodium Lanzaeanum, which last has a very wide range in time; Monocotyledons, by a Sabal and a feather-palm, as well as by the two aquatic genera above mentioned; Gymnosperms, by the extinct araucarian genus Doliostrobus, by rare pine-cones, and by Athrotaxis. Dicotyledonous leaves are not plentiful, the genera recorded being Andromeda, Cinnamomum, Zizyphus, Rhus, Viburnum.

The lignite deposits and pipe-clays of Bovey Tracey in Devon, referred by Heer and Pengelly to the Miocene period, were considered by Gardner to be of the same age as the Bournemouth beds (Middle Eocene). Recent researches show, however, that Heer's view was more nearly correct. The flora of Bovey is like that of the lignite of the Wetterau, which is either highest Oligocene or lowest Miocene. Several species of Nyssa are common to the two districts, as are a climbing palm, two vines, a magnolia, &c. The common tree at Bovey is Sequoia Couttsiae, which probably grew in profusion in the sheltered valleys of Dartmoor, close to the lake. Above these strata in Great Britain there is a complete break, no species of plant ranging upwards into the next fossiliferous division.

Space will not allow us to deal with the numerous scattered deposits which have yielded Tertiary plants. It will be more to the purpose to take distant areas, where the order of the strata is clear, and compare the succession of the floras with that met with in other geographical regions and in other latitudes. For this study it will be most Central and Southern France. convenient to take next south and central France, for in that area can be found a series of plant-bearing strata in which is preserved a nearly continuous history of the vegetation from Upper Eocene down to Pliocene. The account is taken mainly from the writings of Saporta.

The gypsum-deposit of Upper Eocene date at Aix in Provence commences this series, and is remarkable for the variety and perfect preservation of its organic remains. Among its Gymnosperms are numerous Cupressineae of African affinity belonging to the genera Callitris and Widdringtonia, and a juniper close to one indigenous in Greece. Fan-palms, several species of dragon-tree and a banana, like one living in Abyssinia, represent the more peculiar Monocotyledons. Among the noticeable Dicotyledons are the Myricaceae, Proteaceae, Laurineae, Bombax, the Judas-tree, Acacia, Ailanthus, while the most plentiful forms are the Araliaceae. Willows and poplars, with a few other plants of more temperate regions, are found rarely at Aix, and seemingly point to casual introduction from surrounding mountains. In a general way, spiny plants, with stiff branches and dry and coriaceous leaves, dominate the flora, as they now do in Central Africa, to which region on the whole Saporta considers the flora to be most allied.

The succeeding Oligocene flora appears to be more characterized by a gradual replacement of the Eocene species by allied forms, than by any marked change in the assemblage or in the climatic conditions. It forms a perfectly gradual transition to the still newer Miocene period, the newer species slowly appearing and increasing in number. Saporta considers that in central and southern Europe the alternate dry and moist heat of the Eocene period gave place to a climate more equally and more universally humid, and that these conditions continued without material change into the succeeding Miocene stage. Among the types of vegetation which make their appearance in Europe during the Oligocene period may be mentioned the Conifers Libocedrus salicornioides, several species of Chamaecyparis and Sequoia, Taxodium distichum and Glyptostrobus europaeus. The palms include Sabal haeringiana, S. major and Flabellaria. Among the Myricaceae several species of Comptonia are common. These new-comers are all of American type. Aquatic plants, especially water-lilies, are abundant and varied; the soil-dry Callitris and Widdringtonia become scarce.

Though we do not propose to deal with the other European localities for Eocene and Oligocene plants, there is one district to which attention should be drawn, on account of the exceptional state of preservation of the specimens. On the Baltic shores of Prussia there is found a quantity of amber, containing remains of insects and plants. Plants in Amber. This is derived from strata of Oligocene age, and is particularly valuable because it preserves perfectly various soft parts of the plants, which are usually lost in fossil specimens. The tissues, in fact, are preserved just as they would be in Canada balsam. The amber yields such things as fallen flowers, perfect catkins of oak, pollen grains and fungi. It enables us to determine accurately orders and genera which otherwise are unknown in the fossil state, and it thus aids us in forming a truer idea of the flora of the period than can be formed at any locality where the harder parts alone are recognizable. No doubt this amber flora is still imperfectly known, but it is valuable as giving a good idea of the vegetation, during Oligocene times, of a mixed wood of pine and oak, in which there is a mixture of herbaceous and woody plants, such as would now be found under similar conditions.

The plants of which the floral organs or perfect fruits are preserved include the amber-bearing Pinus succinifera, Smilax, Phoenix, the spike of an aroid, 11 species of oak, 2 of chestnut, a beech, Urticaceae, 2 cinnamons and Trianthera among the Lauraceae, representatives of the Cistaceae, Ternstroemiaceae, Dilleniaceae (3 species of Hibbertia), Geraniaceae (Geranium and Erodium), Oxalidaceae, Acer, Celastraceae, Olacaceae, Pittosporaceae, Ilex (2 species), Euphorbiaceae, Umbelliferae (Chaerophyllum), Saxifragaceae (3 genera), Hamamelidaceae, Rosaceae, Connaraceae, Ericaceae (Andromeda and Clethra), Myrsinaceae (3 species), Rubiaceae, Sambucus (2 species), Santalaceae, Loranthaceae (3 species). We here discover for the first time various living families and genera, but there is still a noticeable absence of many of our most prolific existing groups. Whether this deficiency is accidental or real time will show.

The Miocene flora, which succeeds to that just described, is well represented in Europe; but till recently there has been an unfortunate tendency to refer Tertiary floras of all dates to the Miocene period, unless the geological position of the strata was so clear as obviously to forbid this assignment. Thus plant-beds in the basalt of Scotland and Miocene. Ireland were called Miocene; and in the Arctic regions and in North America even plant-beds of Upper Cretaceous age were referred to the same period. The reason for this was that some of the first Tertiary floras to be examined were certainly Miocene, and, when these plants had been studied, it was considered that somewhat similar assemblages found elsewhere in deposits of doubtful geological age must also be Miocene. For a long time it was not recognized that changes in the marine fauna, on which our geological classification mainly depends, correspond scarcely at all with changes in the land plants. It was not suspected, or the fact was ignored, that the break between Cretaceous and Tertiary—made so conspicuous by striking changes in the aquatic animals—had little or no importance in botanical history. It was not realized that an Upper Cretaceous flora needed critical examination to distinguish it from one of Miocene age, and that the two periods were not characterized by a sweeping change of generic type, such as took place among the marine invertebrates. It may appear absurd to a geologist that any one could mistake a Cretaceous flora for one of Miocene date, since the marine animals are completely different and the differences are striking. In the case of the plants, however, the Tertiary generic types in large part appeared in Upper Cretaceous times. Few or no extinct types are to be found in these older strata—there is nothing among the plants equivalent to the unmistakably extinct Ammonites, Belemnites, and a hundred other groups, and we only meet with constant variations in the same genus or family, these variations having seldom any obvious relation to phylogeny.

The Miocene period is unrepresented by any deposits in Great Britain, unless the Bovey lignite should belong to its earliest stage; we will therefore commence with the best known region—that of central Europe and especially of Switzerland, whence a prolific flora has been collected and described by Oswald Heer. The Miocene lacustrine deposits are contained in a number of silted-up lake-basins, which were successively formed and obliterated during the uprise of the Alps and the continuous folding and bending of the earth’s crust which was so striking a feature of the period. These undulations tended to transform valleys into chains of lakes, into which the plants and animals of the surrounding area fell or were washed. We thus find preserved in the Upper Miocene lacustrine deposits of Switzerland a larger flora than is known from any other period of similar length; in fact, an analysis of its composition suggests that the Miocene flora of Switzerland must have been both larger and more varied than that now living in the same country. The best known locality for the Upper Miocene plants is Oeningen, on the Lake of Constance, where have been collected nearly 500 species of plants, the total number of Miocene plants found in Switzerland being stated to be now over 900. Among the characteristics of this Miocene flora are the large number of families represented, the marked increase in the deciduous-leaved plants, the gradual decrease in the number of palms and of tropical plants, and the replacement of these latter by Mediterranean or North American forms. According to Heer, the tropical forms in the Swiss Miocene agree rather with Asiatic types, while the subtropical and temperate plants are allied to forms now living in the temperate zone in North America. Of the 920 species described by Heer, 114 are Cryptogams and 806 flowering plants. Mosses are extremely rare, Heer only describing 3 species. Vascular Cryptogams still include one or two large horsetails with stems over an inch thick, and also 37 species of Fern, amongst the most interesting of which are 5 species belonging to the climbing Lygodium, a genus now living in Java. The number of Ferns is just equal to that now found in Switzerland. Cycads are only represented by fragments of two species, and this seems to be the last appearance of Cycads in Europe. The Coniferae include no fewer than 94 species of Cupressineae and 17 of Abietineae, including several species of Sequoia. Monocotyledons form one-sixth of the known Miocene flora, 25 of them being grasses and 39 sedges; but most of these need further study, and are very insufficiently characterized. Heer records one species of rice and four of millet. Most of the other Monocotyledons call for little remark, though among them is an Iris, a Bromelia and a ginger. Smilax, as in earlier times, was common. Palms, referred to 11 species, are found, though they seem to have decreased in abundance; of them 7 are fan-palms, the others including Phoenicites—a form allied to the date—and a trailing palm, Calamopsis, allied to the canes and rattans. Among the Dicotyledons, the Leguminosae take the first place with 131 species, including Acacia, Caesalpinia and Cassia, each represented by several forms. The occurrence of 90 species of Amentaceae shows that, as the climate became less tropical, the relative proportion of this group to the total flora increased. Evergreen oaks are a marked characteristic of the period, more than half the Swiss species being allied to living American forms. Fig-trees referred to 17 species occur, all with undivided leathery leaves; one is close to the banyan, another to the indiarubber-tree. The Laurineae were plentiful, and include various true laurels, camphor-trees, cinnamon, Persea and Sassafras. The Proteaceae, according to Heer, are still common, the Australian genera Hakea, Dryandra, Grevillea and Banksia, being represented. Amongst gamopetalous plants several of our largest living families, including Campanulaceae, Labiatae, Solanaceae and Primulaceae, are still missing; and of Boragineae, Scrophularineae, Gentianeae and Caprifoliaceae there are only faint and doubtful indications. The Compositae are represented by isolated fruits of various species. Twining lianas are met with in a species of Bignonia; Umbelliferae Ranunculaceae and Cruciferae, are represented by a few fruits. These families, however, do not appear to have had anything like their present importance in the temperate flora, though, as they are mainly herbaceous plants with fruits of moderate hardness, they may have decayed and left no trace. The American Liriodendron still flourished in Europe. Water-lilies of the genera Nymphaea and Nelumbium occur. Maples were still plentiful, 20 species having been described. Rosaceae are rare, Crataegus, Prunus and Amygdalus, being the only genera recorded. It is obvious that many of these Swiss Miocene plants will need more close study before their specific characters, or even their generic position, can be accepted as thoroughly made out; still, this will not affect the general composition of the flora, with its large proportion of deciduous trees and evergreens, and its noticeable deficiency in many of our largest living families.

From Europe it will be convenient to pass to a distant region of similar latitude, so that we may see to what extent botanical provinces existed in Eocene and Oligocene times. It so happens that the interior of temperate North America is almost the only region outside Europe in which a series of plant-bearing strata give a connected history Tertiary of North America. of these periods, and in which the plants have been collected and studied. It is unfortunately still very difficult to correlate even approximately the strata on the two sides of the Atlantic, and there is great doubt as to what strata belong to each division of the Tertiary period even in different parts of North America. This difficulty will disappear as the strata become better known; but at present each of the silted-up lakes has to be studied separately, for we cannot expect so close a correspondence in their faunas and floras as is found in the more crowded and smaller basins in central Europe.

Perhaps the most striking characteristic of the Tertiary floras of North America, as distinguished from those of Europe, is the greater continuity in their history and greater connexion with the existing flora of the same regions. This difference is readily explained when we remember that in Europe the main barriers which stop migration, such as the Alps and the Mediterranean, run east and west, while in America the only barriers of any importance run north and south. In consequence of this peculiarity, climatic or orographic changes in Europe tend to drive animals and plants into a cul de sac, from which there is no escape; but in America similar climatic waves merely cause the species alternately to retreat and advance. This difficulty in migration is probably the reason why the existing European flora is so poor in large-fruited trees compared with what it was in Miocene times or with the existing flora of North America. In America the contrast between the Eocene forests and those now living is much less striking, and this fact has led to the wrong assumption that the present American flora had its origin in the American continent. Such a conclusion is by no means warranted by the facts, for in Tertiary times, as we have seen, the European flora had a distinctly “American” facies. Therefore the so-called American forms may have originated in the Old World, or more probably, as Saporta suggests, in the polar regions, whence they were driven by the increase of cold southwards into Europe and into America. The American Tertiary flora is so large, and the geology of the deposits is so intricate, that it is out of the question to discuss them more fully within the limits of this article. We may point out, however, that the early Tertiary floras seem to indicate a much closer connexion and a greater community of species than is found between the existing plants of Europe and America. Or, rather, we should perhaps say that ancient floras suggest recent dispersal from the place of origin, and less time in which to vary and become modified by the loss of different groups in the two continents. Geographical provinces are certainly indicated by the Eocene flora of Europe and America, but these are less marked than those now existing.

If we turn to a more isolated region, like Australia, we find a Lower Eocene flora distinctly related to the existing flora of Australia and not to that of other continents. Australasia had then as now a peculiar flora of its own, though the former wide dispersal of the Proteaceae and Myrtaceae, and also the large number of Amentaceae then found Australia. in Australia, make the Eocene plants of Europe and Australia much less unlike than are the present floras.

Within the Arctic circle a large number of Tertiary plants have been collected. These were described by Heer, who referred them to the Miocene period; he recognized, in fact, two periods during which Angiosperms flourished within the Arctic regions, the one Upper Cretaceous, the other Miocene. To this view of the Miocene Arctic Regions. age of the plant-bearing strata in Greenland and Spitsbergen there are serious objections, which we will again refer to when the flora has been described.

The Tertiary flora of Greenland is of great interest, from the extremely high latitude at which the plants flourished, thirty of the species having been collected so far north as lat. 81°. Taking first this most northerly locality, in Grinnell Land, we find the flora to comprise 2 horsetails, 11 Conifers (including the living Pinus Abies), 2 grasses, a sedge, 2 poplars, a willow, 2 birches, 2 hazels, an elm, a Viburnum, a water-lily, and a lime. Such an assemblage at the present day would suggest a latitude quite 25° farther south; but it shows decidedly colder conditions than any of the European Eocene, Oligocene, or Miocene strata. From lat. 78° in Spitsbergen Heer records 136 species of fossil plants. More to the south, at Disco Island in lat. 70°, the Tertiary wood seem to have been principally composed of planes and Sequoias; but a large number of other genera occur, the total number of plants already recorded being 137. From various parts of Greenland they now amount to at least 280. Among the plants from Disco, more than a quarter are also found in the Miocene of central Europe. The plants of Disco include, besides the plane and Sequoia, such warm-temperate trees as Ginkgo, oak, beech, poplar, maple, walnut, lime and magnolia. If these different deposits are contemporaneous, as is not improbable, there is a distinct change in the flora as we move farther from the pole, which suggests that difference of latitude then as now was accompanied by a difference in the flora. But if this process is continuous from latitude to latitude, then we ought not to look for a flora of equivalent age in the warm-temperate Miocene deposits of central Europe, but should rather expect to find that the temperate plants of Greenland were contemporaneous with a tropical flora in central Europe. As Mr Starkie Gardner has pointed out, it does not seem reasonable to assume that the same flora could have ranged then through 40° of latitude; it is more probable that an Eocene temperate flora found in the Arctic regions travelled southwards as the climate became cooler, till it became the Miocene temperate flora of central Europe. Mr Gardner suggests, therefore, that the plant-beds of Greenland and Spitsbergen represent the period of greatest heat, and are therefore wrongly referred to the Miocene. At present the evidence is scarcely sufficient to decide the question, for if this view is right, we ought to find within the Arctic circle truly Arctic floras equivalent to the cool Lower Eocene and Miocene periods; but these have not yet been met with.

A steady decrease of temperature marked the Pliocene period throughout Europe, and gradually brought the climatic conditions into correspondence with those now existing, till towards the end of the period neither climate nor physical geography differed greatly from those now existing. Concurrently with this change, the tropical and extinct forms Pliocene. disappeared, and the flora approached more and more nearly to that now existing in the districts where the fossil plants are found, though in the older deposits, at any rate, the geographical distribution still differed considerably from that now met with. At last, in the latest Pliocene strata (often called “pre-Glacial”) we find a flora consisting almost entirely of existing species belonging to the Palaearctic regions, and nearly all still living in the country where the fossils are found. This flora, however, is associated with a fauna of large mammals, the majority of which are extinct.

The plants of the Older Pliocene period are unknown in Great Britain, and little known throughout Europe except in central France and the Mediterranean region. The forests of central France during this epoch showed, according to Saporta, a singular admixture of living European species, with trees now characteristic of the Canary Isles and of North America. For instance, of the living species found at Meximieux, near Lyons, one is American, eight at least belong to the Canaries (six being characteristic of those islands), two are Asiatic, and ten still live in Europe. Taking into account, however, the closest living allies of the fossil plants, we find about equal affinities with the floras of Europe, America, and Asia. There is also a decided resemblance to the earlier Miocene flora. Among the more interesting plants of this deposit may be mentioned Torreya nucifera, now Japanese; an evergreen oak close to the common Quercus Ilex; Laurus canariensis, Apollonias canariensis, Persea carolinensis, and Ilex canariensis; Daphne pontica (a plant of Asia Minor); a species of box, scarcely differing from the English, and a bamboo. To this epoch, or perhaps to a stage slightly later, and not to the Newer Pliocene period, as is generally supposed, should probably be referred the lignite deposits of the Val d'Arno. This lignite and the accompanying leaf-bearing clays underlie and are apparently older than the strata with Newer Pliocene mammals and mollusca. The only mammal actually associated with the plants appears to be a species of tapir, a genus which in Europe seems to be characteristically Miocene and Older Pliocene. The plants of the Val d'Arno have been described by Ristori; they consist mainly of deciduous trees, a large proportion of which are known Miocene and early Pliocene forms, nearly all of them being extinct. A markedly upland character is given to the flora of this valley through the abundance of pines (9 species) and oaks (16 species) which it contains; but this peculiarity is readily accounted for by the steep slopes of the Apennines, which everywhere surround and dominate the old lake-basin. Among the other noticeable plants may be mentioned Betula (3 species), Alnus (2 species), Carpinus, Fagus (4 species), Salix (4 species), Populus (2 species), Platanus, Liquidambar, Planera, Ulmus (2 species), Ficus (2 species), Persoonia, Laurus (5 species), Persea, Sassafras, Cinnamomum (5 species), Oreodaphne, Diospyros (2 species), Andromeda, Magnolia, Acer (3 species), Sapindus, Celastrus (2 species). Ilex (4 species), Rhamnus (3 species), Juglans (5 species), Carya (2 species), Rhus, Myrtus, Crataegus, Prunus, Cassia (3 species). These plants suggest a colder climate than that indicated by the plants of Meximieux—they might, therefore, be thought to belong to a later period. The difference, however, is probably fully accounted for when we take into consideration the biting winds still felt in spring in the valley of the Arno, and the probable large admixture of plants washed down from the mountains above. Somewhat later Pliocene deposits in the Val d'Arno, as well as the tuffs associated with the Pliocene volcanoes in central France, yield plants of a more familiar type, a considerable proportion of them still living in the Mediterranean region, though some are only now found at distant localities, and others are extinct. The flora, however, is essentially Palaearctic, American and Australian types having disappeared.

A somewhat later Pliocene flora is represented by the plants found at Tegelen, near Venloo, on the borders of the Netherlands and Germany. This deposit is of especial interest for the light it throws on the origin of the existing flora of Britain. The Tegelen plants are mainly north European; but there occur others of central and south Europe, and various exotic and extinct forms, nearly all of which, however, belong to the Palaearctic region, though some may now be confined to widely separated parts of it. For instance, Pterocarya caucasica does not grow nearer than the Caucasus, where it is associated with the wild vine—also found at Tegelen; Magnolia Kobus is confined to the north island of Japan; another species of Magnolia cannot be identified and may be extinct. An extinct water-lily, Euryale limburgensis, belongs to a monotypic genus now confined to Assam and China; an extinct sedge, Dulichium vespiforme, belongs to a genus only living in America, though the only living species once flourished also in Denmark; an extinct species of water-aloe (Stratiotes elegans) makes a third genus, represented only by a single living species, which was evidently better represented in Pliocene times. A large proportion of the plants, however, may still be found living in Holland and Britain; but there is a singular scarcity of Composites, though this order is fairly well represented in British strata of slightly later date.

The latest Pliocene, or pre-Glacial, flora of northern Europe is best known from the Cromer Forest-bed of Norfolk and Suffolk, a fluvio-marine deposit which lies beneath the whole of the Glacial deposits of these counties, and passes downwards into the Crag, many of the animals actually associated with the plants being characteristic Pliocene species which seem immediately afterwards to have been exterminated by the increasing cold. The plants contained in the Cromer Forest-bed, of which about 150 species have now been determined, fall mainly into two groups—the forest-trees, and marsh and aquatic plants. We know little or nothing at present of the upland plants, or of those of dry or chalky soils. Forest trees are well represented; they are, in fact, better known than in any of the later English deposits. We find the living British species of Rhamnus, maple, sloe, hawthorn, apple, white-beam, guelder-rose, cornel, elm, birch, alder, hornbeam, hazel, oak, beech, willow, yew and pine, and also the spruce. This is an assemblage that could not well be found under conditions differing greatly from those now holding in Norfolk; there is an absence of both Arctic and south European plants. The variety of trees shows that the climate was mild and moist. Among the herbaceous plants we find, mingled with a number that still live in Norfolk, Hypecoum procumbens, the water-chestnut (Trapa natans), and Najas minor, none of which is now British.

On the Norfolk coast another thin plant-bed occurs locally above the Forest-bed and immediately beneath the Boulder Clay. This deposit shows no trace of forest-trees, but it is full of remains of Arctic mosses, and of the dwarf willow and birch; in short, it yields the flora now found within the Arctic circle.

The incoming of the Glacial epoch does not appear to have been accompanied by any acclimatization of the plants—the species belonging to temperate Europe were locally exterminated, and Arctic forms took their places. The same Arctic flora reappears in deposits immediately above the highest Boulder Clay, deposits formed after the ice Pleistocene. had passed away. These fossil Arctic plants have now been found as far south as Bovey Tracey in Devonshire, where Pengelly and Heer discovered the bear-berry and dwarf birch; London, where also Betula nana occurs; and at Deuben in Saxony, which lies nearly as far south as lat. 50°, but has yielded to Professor Nathorst’s researches several Arctic species of willow and saxifrage. The cold period, however, was not continuous, for both in Great Britain and on the continent of Europe, as well as in Canada, it was broken by the recurrence of a milder climate and the reappearance of a flora almost identical with that now living in the same regions. This “inter-Glacial” flora, though so like that now found in the district, has interesting peculiarities. In England, for instance, it includes Acer monspessulanum, a southern maple which does not now extend nearer than central Europe, and Cotoneaster Pyracantha; also Najas graminea and N. minor, both southern forms not now native of Britain. Brassenia peltata, a water-lily found in the warmer regions almost throughout the world, except in Europe, occurs abundantly in north Germany, but not in Great Britain. Similar inter-Glacial deposits in Tirol contain leaves of Rhododendron ponticum.

Space will not permit us to enter into any full discussion of the recurrence of Glacial and inter-Glacial periods and the influence they may have had on the flora. It is evident, however, that if climatic alternations, such as those just described, are part of the normal routine that has gone on through all geological periods, and are not merely confined to the latest, then such changes must evidently have had great influence on the evolution and geographical distribution both of species and of floras. Whether this was so is a question still to be decided, for in dealing with extinct floras it is difficult to decide, except in the most general way, to what climatic conditions they point. We seem to find indications of long-period climatic oscillations in Tertiary times, but none of the sudden invasion of an Arctic flora, like that which occurred during more recent times. It should not be forgotten, however, that an Arctic flora is mainly distinguishable from a temperate one by its poverty and dwarfed vegetation, its deciduous leaves and small fruits, rather than by the occurrence of any characteristic genera or families. Careful and long-continued study would therefore be needed before we could say of any extinct dwarfed flora that it included only plants which could withstand Arctic conditions.

Authorities.—H. Conwentz, Monographie der baltischen Bernsteinbäume (Danzig, 1890), Die Flora des Bernsteins, vol. ii. (1886); Sir W. Dawson, Papers on the Cretaceous Plants of British North America, Trans. Roy. Soc. Canada (1883–1896); C. von Ettingshausen, “Die Kreideflora von Niederschöna in Sachsen,” Sitz. k. Akad. Wiss. Wien, math.-nat. Cl., vol. lv., Abth. i. (1867); “Report on . . . Fossil Flora of Sheppy,” Proc. Roy. Soc. xxix. 388 (1879); “Report on . . . Fossil Flora of Alum Bay,” ibid. xxx. 228 (1880); C. von Ettingshausen and J. S. Gardner, “Eocene Flora,” vols. i. and ii., Palaeont. Soc. (1879–1886); W. M. Fontaine, “The Potomac or Younger Mesozoic Flora,” U.S. Geological Survey, Monograph xv. (1889); J. S. Gardner, Flora of Alum Bay, in “Geology of the Isle of Wight,” Mem. Geol. Survey (2nd ed., 1889); H. R. Goeppert and A. Menge, Die Flora des Bernsteins und ihre Beziehungen zur Flora der Tertiärformation und der Gegenwart, vol. i. (Danzig, 1883); O. Heer, Flora tertiaria Helvetiae (3 vols., Winterthur, 1855–1859); Flora fossilis arctica (7 vols., Zürich, 1868–1883), “Beiträge zur Kreideflora,—(1) Flora von Moletein in Mähren,” Neue Denkschr. allgem. schweiz. Gesell. Naturwiss., vol. xxiii. mém. 22 (Zürich, 1869–1872); Primaeval World in Switzerland (2 vols., 1876); F. H. Knowlton, “Catalogue of the Cretaceous and Tertiary Plants of North America,” Bull. U.S. Geol. Survey (No. 152, 1898), “Flora of the Montana Formation,” ibid., No. 163 (1900); Krasser, “Die fossile Kreideflora von Kunstadt in Mähren,” Beit. paleont. Geol. Oesterreich-Ungarns, Bd. v. Hft. 3 (1896); Leo. Lesquereux, “Contributions to the Fossil Flora of the Western Territories,” Rep. U.S. Geol. Survey of the Territories, vols. vi., vii., viii. (1877–1883), “The Flora of the Dakota Group,” U.S. Geological Survey, Monograph xvii. (1891); Meschinelli and Squinabol, Flora tertiaria italica (1892); this book contains a full bibliography relating to the Fossil Flora of Italy; J. S. Newberry, “The Flora of Amboy Clays,” U.S. Geological Survey, Monograph xxvi. (1895); Hosius and von der Marck, “Die Flora der westphälischen Kreideformation,” Palaeontographica, vol. xxvi. (1880), and supplement in ibid. vol. xxxi. (1883); A. G. Nathorst, “Glacialflora in Sachsen, am äussersten Rande des nordischen Diluviums,” Kongl. Vetenskaps-Akad. Forh., p. 519 (1894); Clement Reid, “Pliocene Deposits of Britain,” Mem. Geol. Survey (1890), Origin of the British Flora (1899); C. and E. M. Reid, “The Fossil Flora of Tegelen-sur-Meuse, near Venloo, in the Province of Limburg,” Verh. Kon. Akad. Wetensch. Amsterdam, 2e Sect. Dl. xiii. No. 6 (1907); “On the Pre-Glacial Flora of Britain,” Journ. Linn. Soc. (Botany), xxxviii. 206–227 (1908); G. de Saporta, “Prodrome d’une flore fossile des Travertins anciens de Sézanne,” Mém. soc. géol. France, 2nd series, vol. viii. p. 289 (1868); “Recherches sur les végetaux fossiles de Meximieux,” Archiv. Mus. hist. nat. Lyon, i. 131 (1876); Monde des plantes avant l’apparition de l’homme (1879); “Études sur la végetation du sud-est de la France à l’époque tertiare,” Ann. sci. nat. (1862–1888); Flore fossile du Portugal (Lisbon, 1894); G. de Saporta and A. F. Marion, “Essai sur l’état de la végetation a l’époque des marnes heersiennes de Gelinden,” Mém. cour. acad. roy. belgique, vol. xxxvii. No. 6 (1873), and vol. xli. No. 3 (1878); J. Velenovsky, “Die Flora der böhmischen Kreideformation,” in Beiträge zur Paleontologie Oesterreich-Ungarns und des Orients, vols, ii.–v. (1881–1885); Lester F. Ward, “Synopsis of the Flora of the Laramie Group,” 6th Report U.S. Geological Survey, pp. 399–558 (1885); “The Geographical Distribution of Fossil Plants,” 8th Report U.S. Geological Survey, pp. 663–960 (1889); “The Potomac Formation,” 15th Report U.S. Geological Survey, pp. 307–398 (1895); “Some Analogies in the Lower Cretaceous of Europe and America,” 16th Report U.S. Geological Survey, Pt. I., pp. 462–542 (1896); “The Cretaceous Formation of the Black Hills as indicated by the Fossil Plants,” 19th Report U.S. Geological Survey, Pt. II., pp. 521–946 (1899).  (C. R.) 

---

1. In S. speciosum the leaves in a whorl were of unequal size.
2. Endlicher’s name Dadoxylon is conveniently used for Palaeozoic specimens of the kind in question when nothing beyond the wood structure is known.